[opensuse-factory] out of curiosity - / grow from 10 to 12GB
Hi, I noticed that the stuff on my ext4-root-partition has grown by more than 20% over the last few month and I'm not aware that I added anything new. Usual requirements have been for years around 10GB. So, just out of curiosity, did the system miss some housekeeping or does Tumbleweed simply require more space. cu Peter -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Peter Mc Donough composed on 2017-06-03 21:48 (UTC+0200):
I noticed that the stuff on my ext4-root-partition has grown by more than 20% over the last few month and I'm not aware that I added anything new. Usual requirements have been for years around 10GB.
So, just out of curiosity, did the system miss some housekeeping or does Tumbleweed simply require more space.
Which DE(s) is/are installed? Are your logs being rotated? What's in your /var/log/journal/*? Those can get large, old and useless. Is zypper keeping all your installed packages in cache? How many kernels are installed? Here on host gx62b, Plasma and IceWM are the only installed DEs. 5 kernels remain installed. 95% of 5655815 1k total blocks are in use on EXT3 /, compared to 77% of 5655815 for TDE/IceWM on 42.3 with 2 kernels, 87% of 5655815 for TDE/IceWM on 42.1 with 5 kernels, and 76% of 5655815 for KDE4/IceWM on 13.1 with 4 kernels. TW/Plasma does seem rather bloated here. -- "The wise are known for their understanding, and pleasant words are persuasive." Proverbs 16:21 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Op zaterdag 3 juni 2017 21:48:29 CEST schreef Peter Mc Donough:
Hi,
I noticed that the stuff on my ext4-root-partition has grown by more than 20% over the last few month and I'm not aware that I added anything new. Usual requirements have been for years around 10GB.
So, just out of curiosity, did the system miss some housekeeping or does Tumbleweed simply require more space.
cu Peter
If you used the defaults, your root filesystem is btrfs. Another default is to make a snapshot whenever the system is updated. These are probably causing the growth. -- Gertjan Lettink, a.k.a. Knurpht openSUSE Board Member openSUSE Forums Team -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 06/03/2017 04:59 PM, Knurpht - Gertjan Lettink wrote:
Op zaterdag 3 juni 2017 21:48:29 CEST schreef Peter Mc Donough:
Hi,
I noticed that the stuff on my ext4-root-partition has grown by more than 20% over the last few month and I'm not aware that I added anything new. Usual requirements have been for years around 10GB.
So, just out of curiosity, did the system miss some housekeeping or does Tumbleweed simply require more space.
cu Peter
If you used the defaults, your root filesystem is btrfs. Another default is to make a snapshot whenever the system is updated. These are probably causing the growth.
Is there any kind of snapshot rotation policy, or do they grow until manually purged? Nate -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
* Nate Graham
On 06/03/2017 04:59 PM, Knurpht - Gertjan Lettink wrote:
Op zaterdag 3 juni 2017 21:48:29 CEST schreef Peter Mc Donough:
Hi,
I noticed that the stuff on my ext4-root-partition has grown by more than 20% over the last few month and I'm not aware that I added anything new. Usual requirements have been for years around 10GB.
So, just out of curiosity, did the system miss some housekeeping or does Tumbleweed simply require more space.
cu Peter
If you used the defaults, your root filesystem is btrfs. Another default is to make a snapshot whenever the system is updated. These are probably causing the growth.
Is there any kind of snapshot rotation policy, or do they grow until manually purged?
/etc/snapper/configs/root -- (paka)Patrick Shanahan Plainfield, Indiana, USA @ptilopteri http://en.opensuse.org openSUSE Community Member facebook/ptilopteri Registered Linux User #207535 @ http://linuxcounter.net Photos: http://wahoo.no-ip.org/piwigo paka @ IRCnet freenode -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Knurpht - Gertjan Lettink composed on 2017-06-04 00:59 (UTC+0200):
Peter Mc Donough composed: ...
I noticed that the stuff on my ext4-root-partition has grown by more than 20% over the last few month and I'm not aware that I added anything new. Usual requirements have been for years around 10GB....
If you used the defaults, your root filesystem is btrfs. Another default is to make a snapshot whenever the system is updated. These are probably causing the growth.
"for years around" implies an older installation, as does 10GB, likely before the name "Tumbleweed" supplanted "Factory", and before BTRFS became default. And, he did write EXT4 root. Some people aren't agreeable to every default. I saw nothing in OP to suggest OP used them. I have 0 TW installations out of >20 using BTRFS, and probably as many using EXT3 as EXT4. Might OP have done likewise, and not used the BTRFS default? -- "The wise are known for their understanding, and pleasant words are persuasive." Proverbs 16:21 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Check /tmp and /var/tmp, I believe these are never cleaned out by
default. (Reason given being that someone might place permanent stuff
there?)
Also, as Felix mentioned, it would also be useful to check the size of
/var/log to make sure logs are being handled properly.
On 4 June 2017 at 05:48, Peter Mc Donough
Hi,
I noticed that the stuff on my ext4-root-partition has grown by more than 20% over the last few month and I'm not aware that I added anything new. Usual requirements have been for years around 10GB.
So, just out of curiosity, did the system miss some housekeeping or does Tumbleweed simply require more space.
cu Peter
-- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
-- - Karl Cheng (Qantas94Heavy) -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 03/06/17 08:46 PM, Karl Cheng wrote:
Check /tmp and /var/tmp, I believe these are never cleaned out by default. (Reason given being that someone might place permanent stuff there?)
Good point! The OP did not make it clear if any of those were separate mounted file systems. On my system the are, so if my RootFS suddenly grew by 20% I wound *KNOW* that it was not because of anything in /tmp, /var/tmp, /var/log/ or the entries in /var die to spooling, SMTP out being delayed or similar -- If you keep your mind sufficiently open, people will throw a lot of rubbish into it. --William Orton -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Saturday, 3 June 2017 21:48:29 CEST Peter Mc Donough wrote:
Hi,
I noticed that the stuff on my ext4-root-partition has grown by more than 20% over the last few month and I'm not aware that I added anything new. Usual requirements have been for years around 10GB.
So, just out of curiosity, did the system miss some housekeeping or does Tumbleweed simply require more space.
cu Peter I had issues last year with core dumps being saved (and snapshoted) by default, if anything has started crashing, check /var/lib/systemd/coredump/
-- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 06/03/2017 02:48 PM, Peter Mc Donough wrote:
I noticed that the stuff on my ext4-root-partition has grown by more than 20% over the last few month and I'm not aware that I added anything new. Usual requirements have been for years around 10GB.
I'm still under 10G. Check how many kernels you have: ls -l /boot/vmlinuz* If you have more than 3, then maybe the purge-kernel service (or whatever it is called) has not been enabled. Also, open Yast Software Management. Select the "Package groups" view. Click on "orphaned packages". These are packages that are no longer in any of your enable repos. I do a cleanup there from time to time. Note, however, that if you have installed something directly with an RPM, or from a repo that you have since disabled, that will also show up as orphaned. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Am 04.06.2017 um 13:21 schrieb Neil Rickert:
On 06/03/2017 02:48 PM, Peter Mc Donough wrote:
I noticed that the stuff on my ext4-root-partition has grown by more than 20% over the last few month and I'm not aware that I added anything new. Usual requirements have been for years around 10GB.
Addition: This is fresh "KDE-Tumbleweed" from around Nov. 2016, with one "/" and one "/home" partition, both on ext4.
I'm still under 10G.
That size I know from previous openSuse variants, below you can see that tmp is set to be regularly cleaned by the system.
Check how many kernels you have: ls -l /boot/vmlinuz*
the usual suspects: du -hx /boot 72M /boot du -hx /tmp 80K /tmp du -hx /var/tmp 84K /var/tmp du -hx /var/lib/systemd/coredump 2,0M /var/lib/systemd/coredump du -hx /var/log 314M /var/log du -hx / shows 4,2M /root 1,4M /srv 2,3M /bin 8,5M /sbin 14M /lib64 19M /etc 706M /lib 767M /opt (libreoffice 5.2) 1,1G /var 8,8G /usr 12G / /usr with 8.8G is the frist suspect, 59M /usr/sbin 35M /usr/include 692M /usr/lib 2,5G /usr/lib64 132K /usr/local 522M /usr/bin 1,6G /usr/src (two kernels, 4.11.3-x, 4.11.2-x) 3,5G /usr/share
Also, open Yast Software Management. Select the "Package groups" view. Click on "orphaned packages". These are packages that are no longer in any of your enable repos. I do a cleanup there from time to time.
Nothing "red" there. Could you have a look at the size of the biggest directoryies in /usr cu Peter -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Suprred by this thread, I decided to investigate my own disk usage situation and found that BTRFS subvolumes don't seem to be well handled by the various tools that report disk usage. My / partition is 40G, with the full complement of BTRFS subvolumes that the YaST installer sets up by default. I'm trying to figure out how much space is used: $ > df -hl / Filesystem Size Used Avail Use% Mounted on /dev/nvme0n1p2 40G 22G 19G 54% / But that can't be right... $ sudo du -ch / -d 1 [sudo] password for root: 18M /etc 240G /.snapshots 79M /boot 62M /opt 1.4M /srv 361M /tmp 5.5G /usr 1.8G /var 57G /home 4.0K /dev 0 /proc 0 /sys 1.8M /run 2.7M /bin 615M /lib 13M /lib64 0 /mnt 11M /root 12M /sbin 0 /selinux 0 /media 305G / 305G total Heh, that's even worse. Dolphin agrees with df, but Filelight comes up with its own answer and says that 8.9G is used. What is the canonical source of truth for disk usage on BTRFS volumes with subvolumes? Nate On 06/04/2017 07:29 AM, Peter Mc Donough wrote:
Am 04.06.2017 um 13:21 schrieb Neil Rickert:
On 06/03/2017 02:48 PM, Peter Mc Donough wrote:
I noticed that the stuff on my ext4-root-partition has grown by more than 20% over the last few month and I'm not aware that I added anything new. Usual requirements have been for years around 10GB.
Addition: This is fresh "KDE-Tumbleweed" from around Nov. 2016, with one "/" and one "/home" partition, both on ext4.
I'm still under 10G.
That size I know from previous openSuse variants, below you can see that tmp is set to be regularly cleaned by the system.
Check how many kernels you have: ls -l /boot/vmlinuz*
the usual suspects: du -hx /boot 72M /boot du -hx /tmp 80K /tmp du -hx /var/tmp 84K /var/tmp du -hx /var/lib/systemd/coredump 2,0M /var/lib/systemd/coredump du -hx /var/log 314M /var/log
du -hx / shows
4,2M /root 1,4M /srv 2,3M /bin 8,5M /sbin 14M /lib64 19M /etc 706M /lib 767M /opt (libreoffice 5.2) 1,1G /var 8,8G /usr 12G /
/usr with 8.8G is the frist suspect,
59M /usr/sbin 35M /usr/include 692M /usr/lib 2,5G /usr/lib64 132K /usr/local 522M /usr/bin 1,6G /usr/src (two kernels, 4.11.3-x, 4.11.2-x) 3,5G /usr/share
Also, open Yast Software Management. Select the "Package groups" view. Click on "orphaned packages". These are packages that are no longer in any of your enable repos. I do a cleanup there from time to time.
Nothing "red" there.
Could you have a look at the size of the biggest directoryies in /usr
cu Peter
-- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
* Nate Graham
Suprred by this thread, I decided to investigate my own disk usage situation and found that BTRFS subvolumes don't seem to be well handled by the various tools that report disk usage. My / partition is 40G, with the full complement of BTRFS subvolumes that the YaST installer sets up by default. I'm trying to figure out how much space is used:
$ > df -hl / Filesystem Size Used Avail Use% Mounted on /dev/nvme0n1p2 40G 22G 19G 54% /
But that can't be right...
$ sudo du -ch / -d 1 [sudo] password for root: 18M /etc 240G /.snapshots 79M /boot 62M /opt 1.4M /srv 361M /tmp 5.5G /usr 1.8G /var 57G /home 4.0K /dev 0 /proc 0 /sys 1.8M /run 2.7M /bin 615M /lib 13M /lib64 0 /mnt 11M /root 12M /sbin 0 /selinux 0 /media 305G / 305G total
Heh, that's even worse. Dolphin agrees with df, but Filelight comes up with its own answer and says that 8.9G is used.
What is the canonical source of truth for disk usage on BTRFS volumes with subvolumes?
Nate
On 06/04/2017 07:29 AM, Peter Mc Donough wrote:
Am 04.06.2017 um 13:21 schrieb Neil Rickert:
On 06/03/2017 02:48 PM, Peter Mc Donough wrote:
I noticed that the stuff on my ext4-root-partition has grown by more than 20% over the last few month and I'm not aware that I added anything new. Usual requirements have been for years around 10GB.
Addition: This is fresh "KDE-Tumbleweed" from around Nov. 2016, with one "/" and one "/home" partition, both on ext4.
I'm still under 10G.
That size I know from previous openSuse variants, below you can see that tmp is set to be regularly cleaned by the system.
Check how many kernels you have: ls -l /boot/vmlinuz*
the usual suspects: du -hx /boot 72M /boot du -hx /tmp 80K /tmp du -hx /var/tmp 84K /var/tmp du -hx /var/lib/systemd/coredump 2,0M /var/lib/systemd/coredump du -hx /var/log 314M /var/log
du -hx / shows
4,2M /root 1,4M /srv 2,3M /bin 8,5M /sbin 14M /lib64 19M /etc 706M /lib 767M /opt (libreoffice 5.2) 1,1G /var 8,8G /usr 12G /
/usr with 8.8G is the frist suspect,
59M /usr/sbin 35M /usr/include 692M /usr/lib 2,5G /usr/lib64 132K /usr/local 522M /usr/bin 1,6G /usr/src (two kernels, 4.11.3-x, 4.11.2-x) 3,5G /usr/share
Also, open Yast Software Management. Select the "Package groups" view. Click on "orphaned packages". These are packages that are no longer in any of your enable repos. I do a cleanup there from time to time.
Nothing "red" there.
Could you have a look at the size of the biggest directoryies in /usr
cu Peter
man btrfs btrfs fi sh / really??? top posting and no trimming, SHAME ON YOU! -- (paka)Patrick Shanahan Plainfield, Indiana, USA @ptilopteri http://en.opensuse.org openSUSE Community Member facebook/ptilopteri Registered Linux User #207535 @ http://linuxcounter.net Photos: http://wahoo.no-ip.org/piwigo paka @ IRCnet freenode -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Sunday, 4 June 2017 16:31:03 CEST Patrick Shanahan wrote:
btrfs fi sh /
to get more info (e.g. see data/metadata, unallocated etc) use: sudo btrfs filesystem usage -h / (good to add as an alias) -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Sunday 2017-06-04 16:27, Nate Graham wrote:
$ > df -hl / Filesystem Size Used Avail Use% Mounted on /dev/nvme0n1p2 40G 22G 19G 54% /
But that can't be right... $ sudo du -ch / -d 1 [sudo] password for root: 18M /etc 240G /.snapshots
Heh, that's even worse. Dolphin agrees with df, but Filelight comes up with its own answer and says that 8.9G is used.
What is the canonical source of truth for disk usage on BTRFS volumes with subvolumes?
Which truth would you like? https://btrfs.wiki.kernel.org/index.php/FAQ#Understanding_free_space.2C_usin... -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 04/06/17 09:29 AM, Peter Mc Donough wrote:
Addition: This is fresh "KDE-Tumbleweed" from around Nov. 2016, with one "/" and one "/home" partition, both on ext4.
How big is your disk drive, overall? Tucked under my desk I have my old laptop. Now it serves as my MySQL-server and email archiver, but in days gone by it was my workhorse. It had a mere 90G drive (and still does) that was (and still is) set up with LVM. As it turns out I don't use all of the drive. I never did. However with LVM I made the point of having separate /tmp, /var, /usr/share and /srv. Now I also have a partition dedicated to the SQL database. It made backup much easier. It also avoided some vulnerabilities that could arise from hard-linking to between /tmp and /sbin, and others that could arise from soft-linking, because I mounted /tmp "nosuid,nodev,noexec". Heck, ~Downloads and ~MyDocuments and ~MyMovies and ~MyMusic are mounted that way too. A good, if slightly, paranoid, attitude. YMMV on that. As I've pointed out, if my RootFS grows then I'm *sure* something is wrong. When you have all your non-home stuff on your RootFS a lot of things can happen to alter its size and more, some perfectly legitimate, some a slackness in administration, and some that can be dangerous. Some of that dangerous stuff might even be malicious. I don't think my "precautions" are excessive. They may seems lot to people who don't make use of partitioning or LVM, but with LVM, its easy. Yes there are people who think that LVM is complex. Well Bully! Driving a car is more complex! You have a lot more things to consider and end up doing it for a much longer period of time when driving a car. I mention my old laptop because it has only the 90G drive, Its running 13.2 and isn't going to get LEAP'd since its a i586 machine (see 32-bit vs 64-bit thread). Old as it is, it would still server as a writing/browsing/email engine for everyday use, if it wasn't for its weight! -- An ounce of action is worth a ton of theory. -- Friedrich Engels -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Am 04.06.2017 um 17:11 schrieb Anton Aylward:
On 04/06/17 09:29 AM, Peter Mc Donough wrote:
Addition: This is fresh "KDE-Tumbleweed" from around Nov. 2016, with one "/" and one "/home" partition, both on ext4.
How big is your disk drive, overall?
Really big enough. I have three OS on a 250GB SSD, each with 20GB for / and 10GB for /home plus SSD space for virtual machines, plus 1TB HDD for all that for which an SSD could be useful but is not yet cheap enough to retire the HDD. File throughout is ext4. All data, which is not required by the system is on the HDD. Data-backup is as easy as it could be. Over the years, I have been through all that: different partitions for /boot /tmp / and /, LVM, lately btfs. Usually openSuse seldom needed more than 10 GB for the OS, so 20 GB for / has been enough. I think LVM is usefull if several physical drives are combined. With a single SSD the procedure has changed, you simply take away space from one partition and allocate it to an other. Yast ist king.
Tucked under my desk I have my old laptop. Now it serves as my MySQL-server and email archiver, but in days gone by it was my workhorse.
It had a mere 90G drive (and still does) that was (and still is) set up with LVM. As it turns out I don't use all of the drive. I never did.
However with LVM I made the point of having separate /tmp, /var, /usr/share and /srv. Now I also have a partition dedicated to the SQL database.
I don't understood why that should make a system backup easier apart for a backup for the database which is anyhow on a different partition. I used to clean /tmp before I did backups, with my present setup /tmp and /var/tmp are about 200K, so who cares and all other directories have to be backed up anyway.
It made backup much easier. It also avoided some vulnerabilities that could arise from hard-linking to between /tmp and /sbin, and others that could arise from soft-linking, because I mounted /tmp "nosuid,nodev,noexec". Heck, ~Downloads and ~MyDocuments and ~MyMovies and ~MyMusic are mounted that way too. A good, if slightly, paranoid, attitude. YMMV on that.
My data, as I said, is on the HDD and gets an system-independent backup each day. Very nice, I can connect and access it from any OS.
As I've pointed out, if my RootFS grows then I'm *sure* something is wrong.
I also keep an eye on root and I noticed the increase. I like to find out what has grown.
When you have all your non-home stuff on your RootFS a lot of things can happen to alter its size and more, some perfectly legitimate, some a slackness in administration, and some that can be dangerous. Some of that dangerous stuff might even be malicious. ... Yes there are people who think that LVM is complex. Well Bully! Driving a car is more complex! You have a lot more things to consider and end up doing it for a much longer period of time when driving a car.
You are right, driving a car is much more complex than any computer. But LVM would solve a problem I don't have.
I mention my old laptop because it has only the 90G drive, Its running 13.2 and isn't going to get LEAP'd since its a i586 machine (see 32-bit vs 64-bit thread). Old as it is, it would still server as a writing/browsing/email engine for everyday use, if it wasn't for its weight!
What about moving to Debian Jessie, much simpler than driving a car;-) I did it with my little eeePC (1GB RAM, 12GB SSD-Space) works like a charm and gets security updates till 2022, I think. cu Peter -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 04/06/17 01:03 PM, Peter Mc Donough wrote:
How big is your disk drive, overall?
Really big enough. I have three OS on a 250GB SSD, each with 20GB for / and 10GB for /home plus SSD space for virtual machines, plus 1TB HDD for all that for which an SSD could be useful but is not yet cheap enough to retire the HDD.
So you've got a lot of rust. But what part of this pertains to the system under discussion? By deduction, it can't be the one of the three on the 250GB SSD since you say they have 20GB for /. Your /home seems starved. I'm not sure that the "all that" in the 1T HHD means. So where is this 10G / ? or do you mean that its actually 10G used on a 20G allocated? Please show a "df -h" rather than a "du"
I think LVM is usefull if several physical drives are combined.
Well, yes, it is that, but I use if even on single drive machines for "deferred design", that is, I don't need to make hard and fast space allocation provisioning up front. or correcting an imbalance: too much given to one FS and not enough to another. Or a Thin Pool :-)
With a single SSD the procedure has changed, you simply take away space from one partition and allocate it to an other.
Well, that's one way of describing what you do with LVM :-)
And with LVM you can do it with a running system and with mounted file systems
that are in active use.
So how has the procedure changed with SSD? You have to stop the machine,
repartition and then reboot type of changed?
--
Do you want the truth, or a well-designed machination brought
into existence solely for the stroking of your ego?
-- Empty
Am 05.06.2017 um 05:13 schrieb Anton Aylward:
On 04/06/17 01:03 PM, Peter Mc Donough wrote:
How big is your disk drive, overall?
Really big enough. I have three OS on a 250GB SSD, each with 20GB for / and 10GB for /home plus SSD space for virtual machines, plus 1TB HDD for all that for which an SSD could be useful but is not yet cheap enough to retire the HDD.
So you've got a lot of rust. But what part of this pertains to the system under discussion?
By deduction, it can't be the one of the three on the 250GB SSD since you say they have 20GB for /. Your /home seems starved.
No, the "secret" is a second "home" on the HDD for data which is mostly not dependent on any distribution (music, photos, documents) All that is left on /home are just the files which are specific to that OS, everything else is linked to the HDD. An example: I write this on Xubuntu, of its 20GB about 41% are used for / and just 9% of 10GB for /home with 2 users. Tumbleweed uses 62% of its 20GB for / and for /home 18% of its 10GB, again with 2 users, more occupied space because I use with Tumbleweed more programmes. The Suse / was usually around 50%.
or do you mean that its actually 10G used on a 20G allocated? Please show a "df -h" rather than a "du"
Its my pleasure: root@xbu-lux:/home/peter# df -h /vol/openSuse/root_sda1/ Dateisystem Größe Benutzt Verf. Verw% Eingehängt auf /dev/sda1 20G 12G 7,2G 62% /vol/openSuse/root_sda1 and for good measure (thank you Berhard) root@xbu-lux:/home/peter# du -t +1G -xHh /vol/openSuse/root_sda1/ 1,1G /vol/openSuse/root_sda1/var 2,5G /vol/openSuse/root_sda1/usr/lib64 1,6G /vol/openSuse/root_sda1/usr/src 1,1G /vol/openSuse/root_sda1/usr/share/texmf 3,5G /vol/openSuse/root_sda1/usr/share 8,8G /vol/openSuse/root_sda1/usr 12G /vol/openSuse/root_sda1/ And I notice, what ist .../share/texmf? I don't use TeX, is that regularly installed?
Well, yes, it is that, but I use if even on single drive machines for "deferred design", that is, I don't need to make hard and fast space allocation provisioning up front. or correcting an imbalance: too much given to one FS and not enough to another. Or a Thin Pool :-)
Good, but the question is, do you really need different system-partitions or is is just because you are used to it?
With a single SSD the procedure has changed, you simply take away space from one partition and allocate it to an other. Well, that's one way of describing what you do with LVM :-) And with LVM you can do it with a running system and with mounted file systems that are in active use. So how has the procedure changed with SSD? You have to stop the machine, repartition and then reboot type of changed?
True, that is really an important difference. cu Peter -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 05/06/17 06:12 AM, Peter Mc Donough wrote:
And I notice, what ist .../share/texmf? I don't use TeX, is that regularly installed?
Dunno. I don't have that. The way to answer your question is to use 'rpm -q -f' on some of the files in that directory. -- "Is Penetration Testing Worth it? There are two reasons why you might want to conduct a penetration test. One, you want to know whether a certain vulnerability is present because you're going to fix it if it is. And two, you need a big, scary report to persuade your boss to spend more money. If neither is true, I'm going to save you a lot of money by giving you this free penetration test: You're vulnerable. Now, go do something useful about it." -- Bruce Schneier http://www.schneier.com/blog/archives/2007/05/is_penetration.html -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 05/06/17 06:12 AM, Peter Mc Donough wrote:
Well, yes, it is that, but I use if even on single drive machines for "deferred design", that is, I don't need to make hard and fast space allocation provisioning up front. or correcting an imbalance: too much given to one FS and not enough to another. Or a Thin Pool :-)
Good, but the question is, do you really need different system-partitions or is is just because you are used to it?
Are you an Archetypal-Irish to answer a question with another question? I don't know what you mean by "different system-partitions", and you still haven't begun to explain how or why SSD partition management or space management is superior to LVM? There are many justifications to having 'containers': security, backup, conceptual boundaries, packaging. I've discussed many of my personal reasons for LVM and for the size partitions & pools I use many times on this forum, and many people have bloged, written academic articles and given presentations back and forth over this. To paraphrase someone from the security arena, there are two philosophies for security/robustness: one is to compartmentalise, replicate and work on the assumption that any single loss is not crippling; the other is to 'put all your eggs in one basket' and watch the basket very carefully. I'm a proponent of the former. -- There is no legitimate religion apart from truth. --John Calvin -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Am 05.06.2017 um 14:41 schrieb Anton Aylward:
On 05/06/17 06:12 AM, Peter Mc Donough wrote:
Well, yes, it is that, but I use if even on single drive machines for "deferred design", that is, I don't need to make hard and fast space allocation provisioning up front. or correcting an imbalance: too much given to one FS and not enough to another. Or a Thin Pool :-)
Good, but the question is, do you really need different system-partitions or is is just because you are used to it?
Are you an Archetypal-Irish to answer a question with another question? No, just wondering whether LVM for a system is really necessary unless you now that system space requirement are really increasing that much over time.
I don't know what you mean by "different system-partitions", and you still haven't begun to explain how or why SSD partition management or space management is superior to LVM?
Sorry, I look at it from my perspective. Years ago, when I started with Linux, SuSe in fact, it was customary to have several system-partitions on an HDD, for example for / /swap /boot /tmp /usr /var and /home anyway, the reasons for it, I forgot. It was probably the size of HDDs, reliability and one may have needed more than one HDD for a decent system. I don't think it is necessary for what I do with computers to have more than / and /home for each physical OS, especially as my computer is set to BIOS-mode. With UEFI it may be different, but not my problem, yet. As for SSD vs. LVM. LVM is, as you said, superior if you need the flexibility of resizing partitions on a running system the and capability of adding storage space to it. I don't need it presently. It would probably be different if I worked extensively with virtual plus physical machines and really needed the space. I haven't used LV lately, but if I remember correctly, no matter what you do with LV, for booting an OS you need a physical boot partition, which can luckily be in an extended partition but not in a LV.
There are many justifications to having 'containers': security, backup, conceptual boundaries, packaging. I've discussed many of my personal reasons for LVM and for the size partitions & pools I use many times on this forum, and many people have bloged, written academic articles and given presentations back and forth over this.
To paraphrase someone from the security arena, there are two philosophies for security/robustness: one is to compartmentalise, replicate and work on the assumption that any single loss is not crippling; the other is to 'put all your eggs in one basket' and watch the basket very carefully.
I'm a proponent of the former.
The former philosophy requires separation of data from system and different physical drives for either are a must. The latter needs several backup baskets. I run a combination of both. cu Peter -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 05/06/17 12:56 PM, Peter Mc Donough wrote:
No, just wondering whether LVM for a system is really necessary unless you now that system space requirement are really increasing that much over time.
I think you have a dramatic misunderstanding of what LVM is about. leaving aside the matter of Thin Pools, , leaving aside the issue of file systems who seize cannot be adjusted or can only grow, you seem to miss out on my point about "deferred design". If one is experienced, the one will know pretty well how large (or small) to make a partition. You mention that you expect a RootFS to be be about 10G for a basic install. So why make a partition any bigger? Certainly in the days before terabit drives were under $100 space was more critical. The BtrFS approach, "One FS to Rule Them All", one FS for no matter how many spindles/media is a solution. Its all one pool, but the subvolumes offer 'management'. really, if one runs BtrFS, then what's the point in having a seperate /home¿ Well, that's one solution. But back in Ancient Of Days this kind of thing was discouraged since in user space a DoS script could do great damage: while true do mkdir whatever cd whatever done So, yes, partitions with hard boundaries. In a former life I administered a a system with about 150T of database space, RAID_LVM. The spindles were, mostly, 350G. That probably tells you the era :-) The issue here wasn't that LVM could make a file system over multiple "logical" drives, but it was about having a dozen (perhaps, I don't recall) logical database volumes. The volumes needed to be demand-balanced, and that involved moving the LVs to the optimum layout across all the drives. The ability to use 'lvmove' on a live system and watch the metrics made this work. Then there was the matter of moving LVs off a dying 'drive'. What? you ask, didn't Anton say this was RAID? Yes, it was RAID for striping - that is !speed! - not RAID for reliability. Oh, the OS & user space was on a separate subsystem and that was simply mirrored. You can do a lot of things with LVM beyond simply increasing the size of a LV and the file system on it. -- "To ask the right question is already half the solution of a problem". -- Carl Jung. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 05/06/17 03:20 PM, Anton Aylward wrote:
On 05/06/17 12:56 PM, Peter Mc Donough wrote:
No, just wondering whether LVM for a system is really necessary unless you now that system space requirement are really increasing that much over time.
I think you have a dramatic misunderstanding of what LVM is about.
leaving aside the matter of Thin Pools, , leaving aside the issue of file systems who seize cannot be adjusted or can only grow, you seem to miss out on my point about "deferred design".
If one is experienced, the one will know pretty well how large (or small) to make a partition. You mention that you expect a RootFS to be be about 10G for a basic install. So why make a partition any bigger? Certainly in the days before terabit drives were under $100 space was more critical.
The BtrFS approach, "One FS to Rule Them All", one FS for no matter how many spindles/media is a solution. Its all one pool, but the subvolumes offer 'management'. really, if one runs BtrFS, then what's the point in having a seperate /home¿
Well, that's one solution. But back in Ancient Of Days this kind of thing was discouraged since in user space a DoS script could do great damage:
while true do mkdir whatever cd whatever done
So, yes, partitions with hard boundaries.
In a former life I administered a a system with about 150T of database space, RAID_LVM. The spindles were, mostly, 350G. That probably tells you the era :-) The issue here wasn't that LVM could make a file system over multiple "logical" drives, but it was about having a dozen (perhaps, I don't recall) logical database volumes. The volumes needed to be demand-balanced, and that involved moving the LVs to the optimum layout across all the drives.
The ability to use 'lvmove' on a live system and watch the metrics made this work.
Then there was the matter of moving LVs off a dying 'drive'. What? you ask, didn't Anton say this was RAID? Yes, it was RAID for striping - that is !speed! - not RAID for reliability.
Oh, the OS & user space was on a separate subsystem and that was simply mirrored.
You can do a lot of things with LVM beyond simply increasing the size of a LV and the file system on it.
I assign 70GB for my Crucial MX300 525 GB SSD. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Am 05.06.2017 um 21:20 schrieb Anton Aylward:
On 05/06/17 12:56 PM, Peter Mc Donough wrote:
No, just wondering whether LVM for a system is really necessary unless you now that system space requirement are really increasing that much over time.
I think you have a dramatic misunderstanding of what LVM is about.
Probably, my system doesn't require it. You mentioned a Laptop with a 90G HDD with LV. Why would you need an LVM there. It can't be data-protection, you would need a second physical drive for it. As for server-OS space, how much, 5GB?, it is very unlikely that it will suddenly grow beyond 10GB. The database or the mail-archive or the /tmp directory may, in this case wouldn't it make more sense to put everything on one big drive and have something like a full disk alert warning. This problem is so old, I could bet that a solution is probably available for free on each Linux. ...
If one is experienced, the one will know pretty well how large (or small) to make a partition. You mention that you expect a RootFS to be be about 10G for a basic install. So why make a partition any bigger? Certainly in the days before terabit drives were under $100 space was more critical.
Having a nearly empty partition, which may some unknown time in future be required is OK, but wasting disk space when I'm fairly certain I will not need is simply uncool.
The BtrFS approach, "One FS to Rule Them All", one FS for no matter how many spindles/media is a solution. Its all one pool, but the subvolumes offer 'management'. really, if one runs BtrFS, then what's the point in having a seperate /home¿ Well, that's one solution.
btfs has its advantages, but I can't see any for my computer usage. It is very unlikely that I would kept the data to roll back Tumbleweed to solve my original problem, the increase on / from 10G to 12G over three month. If necessary, it will be quicker to install a fresh Tumbleweed.
But back in Ancient Of Days this kind of thing was discouraged since in user space a DoS script could do great damage: .... So, yes, partitions with hard boundaries. In a former life I administered a a system with about 150T of database space, RAID_LVM. The spindles were, mostly, 350G. That probably tells you the era :-) The issue here wasn't that LVM could make a file system over multiple "logical" drives, but it was about having a dozen (perhaps, I don't recall) logical database volumes. The volumes needed to be demand-balanced, and that involved moving the LVs to the optimum layout across all the drives.
This gave me the idea that you use LVM because you are used to it and not because it is necessary. Which is absolutely fine. There are simply more interesting ways of spending our lifetime than changing a partition-layout if there is not obvious advantage in it.
... You can do a lot of things with LVM beyond simply increasing the size of a LV and the file system on it.
But do I need it? Have a nice evening. Peter -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 2017-06-05 22:25, Peter Mc Donough wrote:
Am 05.06.2017 um 21:20 schrieb Anton Aylward:
On 05/06/17 12:56 PM, Peter Mc Donough wrote:
No, just wondering whether LVM for a system is really necessary unless you now that system space requirement are really increasing that much over time.
I think you have a dramatic misunderstanding of what LVM is about.
Probably, my system doesn't require it. You mentioned a Laptop with a 90G HDD with LV. Why would you need an LVM there. It can't be data-protection, you would need a second physical drive for it.
Having a single partition for everything is the thing to do for small systems. Yes, you can tune directories for optimal use, but... it is not that useful.
btfs has its advantages, but I can't see any for my computer usage. It is very unlikely that I would kept the data to roll back Tumbleweed to solve my original problem, the increase on / from 10G to 12G over three month. If necessary, it will be quicker to install a fresh Tumbleweed.
btrfs makes size problems worse. It grows! It is a feature :-) -- Cheers / Saludos, Carlos E. R. (from 42.2 x86_64 "Malachite" at Telcontar)
On 06/05/2017 11:56 AM, Peter Mc Donough wrote:
I haven't used LV lately, but if I remember correctly, no matter what you do with LV, for booting an OS you need a physical boot partition, which can luckily be in an extended partition but not in a LV.
I don't think that's correct. I'm using an LVM because of crypto. So I have an encrypted LVM. I boot Tumbleweed in the encrypted LVM. And "/boot" is part of that encrypted LVM. This is UEFI. There's some boot code in the EFI partition that knows how to prompt for encryption key and access the encrypted LVM to get to "/boot". I think that also works for MBR booting, as long as grub is installed in the MBR and there is enough free space between the MBR and the first partition. The downside of having "/boot" part of the encrypted LVM, is that I have to give the encryption key twice -- first time for grub2, and second time for the kernel. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Am 05.06.2017 um 22:10 schrieb Neil Rickert:
On 06/05/2017 11:56 AM, Peter Mc Donough wrote:
I haven't used LV lately, but if I remember correctly, no matter what you do with LV, for booting an OS you need a physical boot partition, which can luckily be in an extended partition but not in a LV.
I don't think that's correct. I'm using an LVM because of crypto. So I have an encrypted LVM. I boot Tumbleweed in the encrypted LVM. And "/boot" is part of that encrypted LVM. This is UEFI. ...
If you are right, then you are right, Obviously I'm not up-to-date. My computer still uses BIOS-mode, maybe that will change with my next computer? Then I will have to think about it. cu Peter -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
The downside of having "/boot" part of the encrypted LVM, is that I have to give the encryption key twice -- first time for grub2, and second time for the kernel.
This is not necessary. Following http://www.pavelkogan.com/2014/05/23/luks-full-disk-encryption/ http://www.pavelkogan.com/2015/01/25/linux-mint-encryption/ , I created a file keyfile `/crypto_keyfile.bin'. $ dd bs=512 count=4 if=/dev/urandom of=/crypto_keyfile.bin $ cryptsetup luksAddKey /dev/sda1 /crypto_keyfile.bin $ chmod 000 /crypto_keyfile.bin $ chmod -R g-rwx,o-rwx /boot Then I created a file `/etc/crypttab', which contains <dev mapper ID> UUID=<...> /crypto_keyfile.bin Dracut needs the additional file `/etc/dracut.conf.d/99-initcrypt.conf', containing install_items="/crypto_keyfile.bin" Then call `dracut' to regenerate the initramfs image. At least for me, this works like a charm. Werner -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 05/06/17 12:56 PM, Peter Mc Donough wrote:
Years ago, when I started with Linux, SuSe in fact, it was customary to have several system-partitions on an HDD, for example for / /swap /boot /tmp /usr /var and /home anyway, the reasons for it, I forgot. It was probably the size of HDDs, reliability and one may have needed more than one HDD for a decent system.
I don't know about "customary", and I don't know about "reliability". Yes, back in the 1970s when a RK02 was 5M, having a separate drive for /usr (which is where the user accounts lived in the days before USG introduced /home) was pretty much a necessity. By the time we had UNIX (think: "SCO") on a PC drives were a good deal larger and was running the IS, user accounts and application built with a Progress RDBMS on a single drive on a PC. I made good money doing that back in the 1985/87 timeframe. Reliability, then, came from backups, not just of the software, the database, but having a backup machine and parts in the closet. In due course we came to have reliability though better products and though RAID. I think that approach still holds. Compartmentalization is quite another matter. Sometimes it simplifies matters such as backup and restore. More often than not it is about some aspect of reliability. I gave the example in another post of a simple DoS script. The principle holds for programming errors as well as for malice. One of the classical reasons for having separate /tmp and /var and /srv has to do with security. I've mentioned the idea of mounting "nosuid,nodev,noexec". Why? A malicious user can create a (possibly) symbolic link to a file not otherwise accessible to him or her in a directory such as /tmp. If you don't have a seperate /tmp you can't mount it "nosuid,nodev,noexec". The same logic, quite obviously applies to a malicious user breaking in via a network service to /srv. These and other can also apply in /usr/tmp or if the user has an account, under /home. Of course in a shared service environment there may be other reasons that have to do with administration. LVM Thin Pools are one example. I'm sure other users can think of many reasons for compartmentalization even in a single user environment. -- Vizzini: INCONCEIVABLE! Inigo: You keep using that word. I do not think it means what you think it means. -- The Princess Bride -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Something is very wrong here on host gx780. Before updating TW from early April to 20170602, there were 5 installed kernels and 93% of 5.6GB consumed by / on EXT4, with the only installed DEs IceWM and KDE3, no LO, no Gimp, no Java, older (smaller) ESR Firefox, no Chromium no SeaMonkey. Installed are VLC, Kodi, MPV, but the start menu doesn't even have an Office selection, nor any 32bit packages installed except glibc. Before duping I decreased consumption to 84% by reducing installed kernels to 3, then did the dup. After dup, consumption rose to 96% with 4 installed kernels. I forced a log rotate, emptied caches & tempdirs, and removed all journals except current and 1st previous, but / consumption remains 96% with total installed package count at 1589. Most of my TW installations on 5.6GB partitions with only DE KDE3 are less than 90% consumed with 5 kernels installed, and more commonly less than 85%. :~( -- "The wise are known for their understanding, and pleasant words are persuasive." Proverbs 16:21 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Felix Miata composed on 2017-06-05 21:13 (UTC-0400):
Something is very wrong here on host gx780. Before updating TW from early April to 20170602, there were 5 installed kernels and 93% of 5.6GB consumed by / on EXT4, with the only installed DEs IceWM and KDE3, no LO, no Gimp, no Java, older (smaller) ESR Firefox, no Chromium no SeaMonkey. Installed are VLC, Kodi, MPV, but the start menu doesn't even have an Office selection, nor any 32bit packages installed except glibc. Before duping I decreased consumption to 84% by reducing installed kernels to 3, then did the dup. After dup, consumption rose to 96% with 4 installed kernels. I forced a log rotate, emptied caches & tempdirs, and removed all journals except current and 1st previous, but / consumption remains 96% with total installed package count at 1589. Most of my TW installations on 5.6GB partitions with only DE KDE3 are less than 90% consumed with 5 kernels installed, and more commonly less than 85%. :~( Same machine's / consumption of other installations, all on EXT4 except as noted: 13.1 45% of 10.0G KDE4 13.2 52% of 5.6G TDE 42.1 80% of 5.6G TDE 42.2 84% of 5.6G KDE3 Debian 9 72% of 5.6G TDE Kubuntu 16.4.2 71% of 5.6G (EXT3) TDE -- "The wise are known for their understanding, and pleasant words are persuasive." Proverbs 16:21 (New Living Translation)
Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 06/04/2017 08:29 AM, Peter Mc Donough wrote:
Addition: This is fresh "KDE-Tumbleweed" from around Nov. 2016, with one "/" and one "/home" partition, both on ext4.
Okay. I'm now looking at a desktop, at around 10G. It was installed Nov 30, 2016 (I'm going by the date of "/lost+found". It was installed originally as snapshot 20161128 (I'm going by the name of the repo that was the install media). This one is at near 10G. It has KDE, Gnome, MATE, XFCE and LXQt desktops installed. "/home" is part of "/", but is small. It mainly has symlinks to my real home partition mounted at "/xhome". I do that to avoid fights between Tumbleweed and Leap for desktop settings.
From "/usr"
# du -s * 747304 bin 77260 include 788024 lib 3963684 lib64 3316 local 66932 sbin 3312956 share 240128 src 0 tmp 8 X11R6 16 x86_64-suse-linux I do not have kernel sources installed. That might be significant. I think you might be seeing the effect of Parkinson's law for software bloat -- "Software expands to consume all available disk space". -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 06/04/2017 11:08 AM, Neil Rickert wrote:
On 06/04/2017 08:29 AM, Peter Mc Donough wrote:
Addition: This is fresh "KDE-Tumbleweed" from around Nov. 2016, with one "/" and one "/home" partition, both on ext4.
Okay. I'm now looking at a desktop, at around 10G. It was installed Nov 30, 2016 (I'm going by the date of "/lost+found". It was installed originally as snapshot 20161128 (I'm going by the name of the repo that was the install media).
This one is at near 10G. It has KDE, Gnome, MATE, XFCE and LXQt desktops installed.
"/home" is part of "/", but is small. It mainly has symlinks to my real home partition mounted at "/xhome". I do that to avoid fights between Tumbleweed and Leap for desktop settings.
From "/usr"
# du -s * 747304 bin 77260 include 788024 lib 3963684 lib64 3316 local 66932 sbin 3312956 share 240128 src 0 tmp 8 X11R6 16 x86_64-suse-linux
I do not have kernel sources installed. That might be significant.
I think you might be seeing the effect of Parkinson's law for software bloat -- "Software expands to consume all available disk space".
On my system with ext4 partitions. /home and /var have their own partitions: finger@linux-cj1t:~/rtlwifi_sync> sudo du -d 1 -x / [sudo] password for root: 2352 /bin 28428 /etc 2111172 /lib 16 /lost+found 1388 /srv 13576 /lib64 7710532 /usr 4 /opt 4 /mnt 4 /selinux 331484 /boot 8724 /sbin 1316 /tmp 10720 /root 10219724 / The only desktop installed is KDE, and 3 kernels are installed. When I first did the du command, my usage was 11.7 GB. The difference is that /lib/modules had entries for 3 kernels that had been deleted from /boot. That is a place to check. Larry -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 06/04/2017 03:29 PM, Peter Mc Donough wrote:
du -hx / shows
Time for du(1)'s --threshold option to filter stuff greater than a given value (available since coreutils-8.21), e.g.: $ du -t +1G -xh / Have fun, Berny -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Am 04.06.2017 um 21:18 schrieb Bernhard Voelker:
On 06/04/2017 03:29 PM, Peter Mc Donough wrote:
du -hx / shows
Time for du(1)'s --threshold option to filter stuff greater than a given value (available since coreutils-8.21), e.g.:
$ du -t +1G -xh /
Filed, thanks Peter -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 04/06/17 05:18, Peter Mc Donough wrote:
Hi,
I noticed that the stuff on my ext4-root-partition has grown by more than 20% over the last few month and I'm not aware that I added anything new. Usual requirements have been for years around 10GB.
So, just out of curiosity, did the system miss some housekeeping or does Tumbleweed simply require more space.
One thing we did was refactor some of the patterns, which may have caused some new software to be added in some cases but I doubt 2GB worth, something else may have pulled in other dependencies as well. You didn't install some debuginfo or devel packages for any reason? both are capable of taking up alot more system space without much effort. -- Simon Lees (Simotek) http://simotek.net Emergency Update Team keybase.io/simotek SUSE Linux Adelaide Australia, UTC+10:30 GPG Fingerprint: 5B87 DB9D 88DC F606 E489 CEC5 0922 C246 02F0 014B
Am 05.06.2017 um 09:34 schrieb Simon Lees:
On 04/06/17 05:18, Peter Mc Donough wrote:
Hi,
I noticed that the stuff on my ext4-root-partition has grown by more than 20% over the last few month and I'm not aware that I added anything new. Usual requirements have been for years around 10GB.
So, just out of curiosity, did the system miss some housekeeping or does Tumbleweed simply require more space.
One thing we did was refactor some of the patterns, which may have caused some new software to be added in some cases but I doubt 2GB worth, something else may have pulled in other dependencies as well. You didn't install some debuginfo or devel packages for any reason? both are capable of taking up alot more system space without much effort.
Nothing new, except LibreOffice 5.2 instead of 5.3 from Tumbleweed. I want the Libreofffice offline-help. 5.3 is removed as much as possible. So that space is compensated for. But tested from xubuntu I found 1.1G lots of tex-stuff in .../share/texmf. I don't use TeX, is that installed by default? root@xbu-lux:/home/peter# du -t +1G -xHh /vol/openSuse/root_sda1/ 1,1G /vol/openSuse/root_sda1/var 2,5G /vol/openSuse/root_sda1/usr/lib64 1,6G /vol/openSuse/root_sda1/usr/src 1,1G /vol/openSuse/root_sda1/usr/share/texmf 3,5G /vol/openSuse/root_sda1/usr/share 8,8G /vol/openSuse/root_sda1/usr 12G /vol/openSuse/root_sda1/ cu Peter -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Am 03.06.2017 um 21:48 schrieb Peter Mc Donough:
Hi,
I noticed that the stuff on my ext4-root-partition has grown by more than 20% over the last few month and I'm not aware that I added anything new. Usual requirements have been for years around 10GB.
So, just out of curiosity, did the system miss some housekeeping or does Tumbleweed simply require more space.
I think I found what inflated my / system by roughly 2GB. Tons of "texlive" packages have been added. These obviously belong to TeX/LaTeX, which I haven't missed since I use (open)SuSe, i.e. for 15 years. What has changed that these packages are now a requirement for Tumbleweed? cu Peter ps. Spell checking suggested replacing "texlive" by "expletive" -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
* Peter Mc Donough
Am 03.06.2017 um 21:48 schrieb Peter Mc Donough:
Hi,
I noticed that the stuff on my ext4-root-partition has grown by more than 20% over the last few month and I'm not aware that I added anything new. Usual requirements have been for years around 10GB.
So, just out of curiosity, did the system miss some housekeeping or does Tumbleweed simply require more space.
I think I found what inflated my / system by roughly 2GB.
Tons of "texlive" packages have been added. These obviously belong to TeX/LaTeX, which I haven't missed since I use (open)SuSe, i.e. for 15 years.
What has changed that these packages are now a requirement for Tumbleweed?
what did you install that requires tex* rpm -q --whatrequires texlive should answer your question perhaps requirements changed for a package you have installed. rpm -qa *tex* provides 8 packages on my Tw and the only texlive package is a font
ps. Spell checking suggested replacing "texlive" by "expletive"
comment not appropriate, opensoftware is provided by people volunteering their time and service, including those here you want to answer your questions. they do not deserve your derision. -- (paka)Patrick Shanahan Plainfield, Indiana, USA @ptilopteri http://en.opensuse.org openSUSE Community Member facebook/ptilopteri Registered Linux User #207535 @ http://linuxcounter.net Photos: http://wahoo.no-ip.org/piwigo paka @ IRCnet freenode -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Patrick Shanahan composed on 2017-06-08 16:55 (UTC-0400):
Peter Mc Donough composed:
ps. Spell checking suggested replacing "texlive" by "expletive"
comment not appropriate, opensoftware is provided by people volunteering their time and service, including those here you want to answer your questions. they do not deserve your derision. . What derision? All I see is a report of a surprising behavior of his spell checker. -- "The wise are known for their understanding, and pleasant words are persuasive." Proverbs 16:21 (New Living Translation)
Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Am 08.06.2017 um 22:55 schrieb Patrick Shanahan:
* Peter Mc Donough
[06-08-17 16:34]: Am 03.06.2017 um 21:48 schrieb Peter Mc Donough: ... I think I found what inflated my / system by roughly 2GB. Tons of "texlive" packages have been added. ... what did you install that requires tex* rpm -q --whatrequires texlive should answer your question
The following gave 886 lines, as user and as root, I copy the first and the last line. rpm -q --whatrequires texlive | sort > atemp_what_requires.txt texlive-12many-2016.113.0.0.3svn15878-28.2.noarch ... texlive-zapfding-2016.113.svn31835-27.2.noarch
perhaps requirements changed for a package you have installed. rpm -qa *tex* provides 8 packages on my Tw and the only texlive package is a font
on my computer, as root and as user, rpm -qa *tex* no output??? The only thing I changed was installing Libreoffic 5.2. I have removed it now an use the one from Tumbleweed. Could you help me with the appropriate zypper settings for removing texlive. An important one I found already --dry-run.
ps. Spell checking suggested replacing "texlive" by "expletive" comment not appropriate, ... people volunteering ... ... do not deserve your derision.
I'm sorry and you are right, without those volunteers I wouldn't have been able to use SuSe for 15 "expletive" interesting years. cu Peter -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
* Peter Mc Donough
Am 08.06.2017 um 22:55 schrieb Patrick Shanahan:
* Peter Mc Donough
[06-08-17 16:34]: Am 03.06.2017 um 21:48 schrieb Peter Mc Donough: ... I think I found what inflated my / system by roughly 2GB. Tons of "texlive" packages have been added. ... what did you install that requires tex* rpm -q --whatrequires texlive should answer your question
The following gave 886 lines, as user and as root, I copy the first and the last line.
rpm -q --whatrequires texlive | sort > atemp_what_requires.txt
texlive-12many-2016.113.0.0.3svn15878-28.2.noarch ... texlive-zapfding-2016.113.svn31835-27.2.noarch
then rpm -q --whatrequires texlive | grep -v tex
perhaps requirements changed for a package you have installed. rpm -qa *tex* provides 8 packages on my Tw and the only texlive package is a font
on my computer, as root and as user, rpm -qa *tex* no output???
then: rpm -qa |grep -v tex
The only thing I changed was installing Libreoffic 5.2. I have removed it now an use the one from Tumbleweed.
Could you help me with the appropriate zypper settings for removing texlive. An important one I found already --dry-run.
zypper -v rm texlive I don't understand "--dry-run" as I always have to confirm zypper actions unless I use "-y" and I don't. but really, using "SuSe for 15 "expletive" interesting years" should have provided you with the knowledge to figure out these things yourself, unless you count sitting in front of a monitor as using .... expletive deleted! -- (paka)Patrick Shanahan Plainfield, Indiana, USA @ptilopteri http://en.opensuse.org openSUSE Community Member facebook/ptilopteri Registered Linux User #207535 @ http://linuxcounter.net Photos: http://wahoo.no-ip.org/piwigo paka @ IRCnet freenode -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Am 09.06.2017 um 17:40 schrieb Patrick Shanahan:
* Peter Mc Donough
[06-09-17 09:40]: Am 08.06.2017 um 22:55 schrieb Patrick Shanahan:
* Peter Mc Donough
[06-08-17 16:34]: Am 03.06.2017 um 21:48 schrieb Peter Mc Donough: ... on my computer, as root and as user, rpm -qa *tex* no output???
then: rpm -qa |grep -v tex Doesn't help about 2500 lines, with, I think, almost everything the computer uses. ...
Could you help me with the appropriate zypper settings for removing texlive. An important one I found already --dry-run.
zypper -v rm texlive
Executed, after a short look at the output, accepted. Advice zypper ps -s, too many services need a restart. So, reboot, everything seems to work fine an the system is back to about 10GB disk space.
but really, using "SuSe for 15 "expletive" interesting years" should have provided you with the knowledge to figure out these things yourself, unless you count sitting in front of a monitor as using ....
It is your and the other contributors fault, my Tumbleweed runs without too many serious problems;-) and my problem was not one of openSuSe. Thanks Peter -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
* Peter Mc Donough
Am 09.06.2017 um 17:40 schrieb Patrick Shanahan:
* Peter Mc Donough
[06-09-17 09:40]: Am 08.06.2017 um 22:55 schrieb Patrick Shanahan:
* Peter Mc Donough
[06-08-17 16:34]: Am 03.06.2017 um 21:48 schrieb Peter Mc Donough: ... on my computer, as root and as user, rpm -qa *tex* no output???
then: rpm -qa |grep -v tex Doesn't help about 2500 lines, with, I think, almost everything the computer uses.
yeah, wrong from me, s/b: rpm -q --whatrequires texlive |grep -v tex
...
Could you help me with the appropriate zypper settings for removing texlive. An important one I found already --dry-run.
zypper -v rm texlive
Executed, after a short look at the output, accepted. Advice zypper ps -s, too many services need a restart. So, reboot, everything seems to work fine an the system is back to about 10GB disk space.
that's good.
but really, using "SuSe for 15 "expletive" interesting years" should have provided you with the knowledge to figure out these things yourself, unless you count sitting in front of a monitor as using ....
It is your and the other contributors fault, my Tumbleweed runs without too many serious problems;-) and my problem was not one of openSuSe.
sorry I took the time to help you, you will forgive me? and for quite some time now, it is openSUSE -- (paka)Patrick Shanahan Plainfield, Indiana, USA @ptilopteri http://en.opensuse.org openSUSE Community Member facebook/ptilopteri Registered Linux User #207535 @ http://linuxcounter.net Photos: http://wahoo.no-ip.org/piwigo paka @ IRCnet freenode -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Peter Mc Donough [09.06.2017 15:39]:
on my computer, as root and as user, rpm -qa *tex* no output???
If you have a file in the current directory that contains "tex" in its name, the shell expand *tex* properly ;) Better use rpm -qa '*tex*', so that expansion will not occur. Werner --
Am 15.06.2017 um 11:50 schrieb Werner Flamme:
Peter Mc Donough [09.06.2017 15:39]:
on my computer, as root and as user, rpm -qa *tex* no output???
If you have a file in the current directory that contains "tex" in its name, the shell expand *tex* properly ;)
Better use rpm -qa '*tex*', so that expansion will not occur.
I'll keep that in mind. The '*tex*' problem itself is solved. Thank's Peter -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Op donderdag 8 juni 2017 22:32:59 CEST schreef Peter Mc Donough:
Am 03.06.2017 um 21:48 schrieb Peter Mc Donough:
Hi,
I noticed that the stuff on my ext4-root-partition has grown by more than 20% over the last few month and I'm not aware that I added anything new. Usual requirements have been for years around 10GB.
So, just out of curiosity, did the system miss some housekeeping or does Tumbleweed simply require more space.
I think I found what inflated my / system by roughly 2GB.
Tons of "texlive" packages have been added. These obviously belong to TeX/LaTeX, which I haven't missed since I use (open)SuSe, i.e. for 15 years.
I used: zypper addlock texlive to prevent these packages from loading. -- fr.gr. member openSUSE Freek de Kruijf -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
* Freek de Kruijf
Op donderdag 8 juni 2017 22:32:59 CEST schreef Peter Mc Donough:
Am 03.06.2017 um 21:48 schrieb Peter Mc Donough:
Hi,
I noticed that the stuff on my ext4-root-partition has grown by more than 20% over the last few month and I'm not aware that I added anything new. Usual requirements have been for years around 10GB.
So, just out of curiosity, did the system miss some housekeeping or does Tumbleweed simply require more space.
I think I found what inflated my / system by roughly 2GB.
Tons of "texlive" packages have been added. These obviously belong to TeX/LaTeX, which I haven't missed since I use (open)SuSe, i.e. for 15 years.
I used:
zypper addlock texlive
to prevent these packages from loading.
but this presumes you already know that they will be installed and you are certain you do not want/need them. I believe the op is wanting to know why they were installed and he didn't expect the install and now does not want it. he has the method to provide the why but has not, yet. -- (paka)Patrick Shanahan Plainfield, Indiana, USA @ptilopteri http://en.opensuse.org openSUSE Community Member facebook/ptilopteri Registered Linux User #207535 @ http://linuxcounter.net Photos: http://wahoo.no-ip.org/piwigo paka @ IRCnet freenode -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Am 09.06.2017 um 10:08 schrieb Freek de Kruijf:
Op donderdag 8 juni 2017 22:32:59 CEST schreef Peter Mc Donough:
Am 03.06.2017 um 21:48 schrieb Peter Mc Donough:
... I think I found what inflated my / system by roughly 2GB. Tons of "texlive" packages have been added. ... I used: zypper addlock texlive to prevent these packages from loading.
All "texlive" cleared, zypper addlock texlive active. Thanks Peter -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Op zaterdag 10 juni 2017 11:26:49 CEST schreef Peter Mc Donough:
Am 09.06.2017 um 10:08 schrieb Freek de Kruijf:
Op donderdag 8 juni 2017 22:32:59 CEST schreef Peter Mc Donough:
Am 03.06.2017 um 21:48 schrieb Peter Mc Donough:
...
I think I found what inflated my / system by roughly 2GB. Tons of "texlive" packages have been added. ...
I used: zypper addlock texlive to prevent these packages from loading.
All "texlive" cleared, zypper addlock texlive active.
Thanks Peter
Still wondering why texlive gets installed without the lock when there is nothing that requires it. "rpm -q --whatrequires texlive" returns nothing. -- fr.gr. member openSUSE Freek de Kruijf -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Sat, 10 Jun 2017, 12:35:29 +0200, Freek de Kruijf wrote:
Op zaterdag 10 juni 2017 11:26:49 CEST schreef Peter Mc Donough:
Am 09.06.2017 um 10:08 schrieb Freek de Kruijf:
Op donderdag 8 juni 2017 22:32:59 CEST schreef Peter Mc Donough:
Am 03.06.2017 um 21:48 schrieb Peter Mc Donough:
...
I think I found what inflated my / system by roughly 2GB. Tons of "texlive" packages have been added. ...
I used: zypper addlock texlive to prevent these packages from loading.
All "texlive" cleared, zypper addlock texlive active.
Thanks Peter
Still wondering why texlive gets installed without the lock when there is nothing that requires it. "rpm -q --whatrequires texlive" returns nothing.
probably something _recommends_ it, and you didn't specify --no-recommends Cheers. l8er manfred
Op zaterdag 10 juni 2017 12:42:03 CEST schreef Manfred Hollstein:
On Sat, 10 Jun 2017, 12:35:29 +0200, Freek de Kruijf wrote:
Still wondering why texlive gets installed without the lock when there is nothing that requires it. "rpm -q --whatrequires texlive" returns nothing.
probably something _recommends_ it, and you didn't specify --no-recommends
I also tried "rpm -q --whatrecommends texlive": nothing. Also the other what....: nothing. -- fr.gr. member openSUSE Freek de Kruijf -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Op zaterdag 10 juni 2017 12:53:17 CEST schreef Freek de Kruijf:
Op zaterdag 10 juni 2017 12:42:03 CEST schreef Manfred Hollstein:
On Sat, 10 Jun 2017, 12:35:29 +0200, Freek de Kruijf wrote:
Still wondering why texlive gets installed without the lock when there is nothing that requires it. "rpm -q --whatrequires texlive" returns nothing.
probably something _recommends_ it, and you didn't specify --no-recommends
I also tried "rpm -q --whatrecommends texlive": nothing. Also the other what....: nothing.
Removing the lock on texlive and specifying --no-recommends on zypper dup indeed did not pull in somewhat more than 1700 packages. -- fr.gr. member openSUSE Freek de Kruijf -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Jun 10 2017, Freek de Kruijf
Op zaterdag 10 juni 2017 12:42:03 CEST schreef Manfred Hollstein:
On Sat, 10 Jun 2017, 12:35:29 +0200, Freek de Kruijf wrote:
Still wondering why texlive gets installed without the lock when there is nothing that requires it. "rpm -q --whatrequires texlive" returns nothing.
probably something _recommends_ it, and you didn't specify --no-recommends
I also tried "rpm -q --whatrecommends texlive": nothing. Also the other what....: nothing.
That will only give you the dependencies that use the name "texlive". But there are many ways to spell that dependency. Andreas. -- Andreas Schwab, SUSE Labs, schwab@suse.de GPG Key fingerprint = 0196 BAD8 1CE9 1970 F4BE 1748 E4D4 88E3 0EEA B9D7 "And now for something completely different." -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Op maandag 12 juni 2017 09:34:33 CEST schreef Andreas Schwab:
On Jun 10 2017, Freek de Kruijf
wrote: Op zaterdag 10 juni 2017 12:42:03 CEST schreef Manfred Hollstein:
On Sat, 10 Jun 2017, 12:35:29 +0200, Freek de Kruijf wrote:
Still wondering why texlive gets installed without the lock when there is nothing that requires it. "rpm -q --whatrequires texlive" returns nothing.
probably something _recommends_ it, and you didn't specify --no-recommends
I also tried "rpm -q --whatrecommends texlive": nothing. Also the other what....: nothing.
That will only give you the dependencies that use the name "texlive". But there are many ways to spell that dependency.
Andreas.
It is a recommend. When I remove the lock and do zypper without --no- recommends around 1700 packages will be pulled in, with --no-recommends nothing will be pulled in. -- fr.gr. member openSUSE Freek de Kruijf -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Hello, Am Montag, 12. Juni 2017, 10:00:18 CEST schrieb Freek de Kruijf:
It is a recommend. When I remove the lock and do zypper without --no- recommends around 1700 packages will be pulled in, with --no-recommends nothing will be pulled in.
Being able to reproduce it means we are getting closer to finding out the reason ;-) First, please create a solver test case as described on https://en.opensuse.org/SDB:Zypper_troubleshooting Then go to http://gk2.sk/zypper-dependency-graph/ and follow the steps described there to get a nice dependency graph. (The article is quite old, I hope it still works.) And finally, please report back which package has the dependency ;-) Regards, Christian Boltz -- Nööö ... das will ich nicht. Ich bin mitlerweile viel zu bequem geworden und will mir mein KDE nicht mehr wegdenken. (Ausserdem muss ich meine Hardware auch nutzen - sonst hätte ich sie mir nicht kaufen müssen! Wozu 512 MB, wenn ich nicht mal 256 MB voll kriegen würde? Hier hilft KDE sehr! :-)) ) [Konrad Neitzel in suse-linux] -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Op maandag 12 juni 2017 22:03:51 CEST schreef Christian Boltz:
Hello,
Am Montag, 12. Juni 2017, 10:00:18 CEST schrieb Freek de Kruijf:
It is a recommend. When I remove the lock and do zypper without --no- recommends around 1700 packages will be pulled in, with --no-recommends nothing will be pulled in.
Being able to reproduce it means we are getting closer to finding out the reason ;-)
First, please create a solver test case as described on https://en.opensuse.org/SDB:Zypper_troubleshooting
Did that.
Then go to http://gk2.sk/zypper-dependency-graph/ and follow the steps described there to get a nice dependency graph. (The article is quite old, I hope it still works.)
Unfortunately the graph making is no longer supported and the text file did not mention anything about dependencies.
And finally, please report back which package has the dependency ;-)
So nothing to report, except failure. -- fr.gr. member openSUSE Freek de Kruijf -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Hello, Am Dienstag, 13. Juni 2017, 14:07:26 CEST schrieb Freek de Kruijf:
Op maandag 12 juni 2017 22:03:51 CEST schreef Christian Boltz:
Am Montag, 12. Juni 2017, 10:00:18 CEST schrieb Freek de Kruijf:
It is a recommend. When I remove the lock and do zypper without --no- recommends around 1700 packages will be pulled in, with --no-recommends nothing will be pulled in.
Being able to reproduce it means we are getting closer to finding out the reason ;-)
First, please create a solver test case as described on https://en.opensuse.org/SDB:Zypper_troubleshooting
Did that.
Then go to http://gk2.sk/zypper-dependency-graph/ and follow the steps described there to get a nice dependency graph. (The article is quite old, I hope it still works.)
Unfortunately the graph making is no longer supported and the text file did not mention anything about dependencies.
:-( I'm not an expert for solver testcases (actually I looked at one for the first time 5 minutes ago ;-) - but it looks like /var/log/zypper.solverTestCase/solver-system.xml.gz contains all the dependencies (including Recommends) for all packages. Search for "texlive" or just "tex" there ;-) I also just found zypp-NameReqPrv - have a look at its manpage ;-) The boring alternative is to search for a zypp expert and ask him/her to check the testcase ;-) Regards, Christian Boltz -- Es könnte zum Beispiel sein, daß Du inzwischen besser bist als 95% der anderen Teilnehmer hier. Das ist für mindestens 45% der Leute, die das von sich glauben jedoch nicht der Fall. :-) [Kristian Koehntopp in suse-linux] -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Peter Mc Donough wrote:
I think I found what inflated my / system by roughly 2GB.
Tons of "texlive" packages have been added. These obviously belong to TeX/LaTeX, which I haven't missed since I use (open)SuSe, i.e. for 15 years.
What has changed that these packages are now a requirement for Tumbleweed?
They are not, at least not in my case. neither on TW nor on Leap 42.2. Check /var/log/zypp/history, for the first occurence of texlive (other than texlive-lm-fonts which might be installed without texlive) I have it installed on purpose, as I work in Science/Astronomy where LaTeX is standard . But I had to do this 'manually'. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Op vrijdag 9 juni 2017 18:07:57 CEST schreef pit:
Peter Mc Donough wrote:
I think I found what inflated my / system by roughly 2GB.
Tons of "texlive" packages have been added. These obviously belong to TeX/LaTeX, which I haven't missed since I use (open)SuSe, i.e. for 15 years.
What has changed that these packages are now a requirement for Tumbleweed?
They are not, at least not in my case. neither on TW nor on Leap 42.2. Check /var/log/zypp/history, for the first occurence of texlive (other than texlive-lm-fonts which might be installed without texlive)
I have it installed on purpose, as I work in Science/Astronomy where LaTeX is standard . But I had to do this 'manually'.
I don't have it installed knurphtlaptop:/home/install/iso # rpm -qa | grep -i texlive texlive-lm-fonts-2016.113.2.004svn28119-27.2.noarch so the OP must have installed something that pulled those packages either as dependencies or recommends. No other way. -- Gertjan Lettink, a.k.a. Knurpht openSUSE Board Member openSUSE Forums Team -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Am 09.06.2017 um 18:07 schrieb pit:
Peter Mc Donough wrote:
I think I found what inflated my / system by roughly 2GB. Tons of "texlive" packages have been added. These obviously belong to TeX/LaTeX, which I haven't missed since I use (open)SuSe, i.e. for 15 years. What has changed that these packages are now a requirement for Tumbleweed?
They are not, at least not in my case. neither on TW nor on Leap 42.2. Check /var/log/zypp/history, for the first occurence of texlive (other than texlive-lm-fonts which might be installed without texlive)
Bingo, there seems to be the whole set. 2017-05-10 12:20:56|install|texlive-12many-doc| ... 2017-05-10 12:22:56|install|texlive-zapfding-fonts Anyway, I followed Patrick's advise. zypper -v rm texlive Nothing suspicious showed up. so, klick and done. Diskspace is back to 10GB Thanks Peter -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Peter Mc Donough wrote:
Am 09.06.2017 um 18:07 schrieb pit:
Peter Mc Donough wrote:
I think I found what inflated my / system by roughly 2GB. Tons of "texlive" packages have been added. These obviously belong to TeX/LaTeX, which I haven't missed since I use (open)SuSe, i.e. for 15 years. What has changed that these packages are now a requirement for Tumbleweed?
They are not, at least not in my case. neither on TW nor on Leap 42.2. Check /var/log/zypp/history, for the first occurence of texlive (other than texlive-lm-fonts which might be installed without texlive)
Bingo, there seems to be the whole set. 2017-05-10 12:20:56|install|texlive-12many-doc| ... 2017-05-10 12:22:56|install|texlive-zapfding-fonts
The interesting part would be the 'command' line that belongs to this install, something like 2017-05-10 12:22:56|command|root@linux-84f0|'zypper' 'install' ..... it should tell you what actually triggered texlive. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Am 10.06.2017 um 14:24 schrieb pit:
Peter Mc Donough wrote:
Am 09.06.2017 um 18:07 schrieb pit:
Peter Mc Donough wrote:
... They are not, at least not in my case. neither on TW nor on Leap 42.2. Check /var/log/zypp/history, for the first occurence of texlive (other than texlive-lm-fonts which might be installed without texlive)
Bingo, there seems to be the whole set. 2017-05-10 12:20:56|install|texlive-12many-doc| ... 2017-05-10 12:22:56|install|texlive-zapfding-fonts
The interesting part would be the 'command' line that belongs to this install, something like 2017-05-10 12:22:56|command|root@linux-84f0|'zypper' 'install' ..... it should tell you what actually triggered texlive.
Below are the 9 lines which precede 2017-05-10 texlive Installed got somthing with "perl" which doesn't look suspicious. I can only assume, that I added somthing before 2017-10-08, which brough in texlive on 2017-05-10 Anyhow, I got rid of all "textlive", marked it "taboo" in Yast, unticked the checkbox "recommnds" in Yast and added solver.onlyRequires = true in /etc/zypp/zypp.conf. This should avoid all accidential downloads in future. Thanks Peter ------------ previous zypper dup --no-allow-vendor-change ... 2017-05-08 09:11:28|install|samba-winbind|4.6.3+git.21.0735c828d4f-1.1|x86_64||repo-oss|afc8a09bd632e23af1dcd134ffbe4b43f01f8dd1| # 2017-05-08 09:12:18 Output of open-iscsi-2.0.874-33.1.x86_64.rpm %posttrans script: # Creating initrd: /boot/initrd-4.10.12-1-default # dracut: Executing: /usr/bin/dracut --logfile /var/log/YaST2/mkinitrd.log --force /boot/initrd-4.10.12-1-default 4.10.12-1-default # dracut: dracut module 'systemd-bootchart' will not be installed, because command '/usr/lib/systemd/systemd-bootchart' could not be found! # dracut: *** Including module: bash *** ... lots of dracut # dracut: *** Creating initramfs image file '/boot/initrd-4.10.13-1-default' done *** next zypper dup --no-allow-vendor-change 2017-05-10 12:20:55|command|root@lux.site|'/usr/bin/ruby' '/usr/lib/YaST2/bin/y2start' 'sw_single' 'qt'| 2017-05-10 12:20:55|install|perl-CGI|4.36-1.1|noarch||repo-oss|ed02f13c1c21f20556e09adaa8dd831d5725af4b| 2017-05-10 12:20:55|install|perl-File-Copy-Recursive|0.38-9.4|noarch||repo-oss|0f89cc25f14b6ff39ea40512740555ebf5ab0558| 2017-05-10 12:20:55|install|perl-File-Which|1.21-1.1|noarch||repo-oss|2b0f0ae810f65cd9bf19299280b4df3d49a4e62a| 2017-05-10 12:20:55|install|perl-HTML-Form|6.03-5.5|noarch||repo-oss|37f7f8962dfe0fbd734a20352cb5654b60ad7d05| 2017-05-10 12:20:55|install|perl-HTML-Tree|5.03-1.4|noarch||repo-oss|24caa71aae0b0cde9e4e05ad406f00c9e0997bc9| 2017-05-10 12:20:55|install|perl-IPC-System-Simple|1.25-3.4|noarch||repo-oss|ad937e57f4935f024f31a0691cf577d7786ac9ab| 2017-05-10 12:20:56|install|perl-Sub-Uplevel|0.240.0-13.4|noarch||repo-oss|66f42be36bf4cb402d5cc7f3634a8a894923ff1f| 2017-05-10 12:20:56|install|perl-YAML-Tiny|1.70-1.1|noarch||repo-oss|c8a9bc190c680beb5f1658414fb791e17d41a9ad| 2017-05-10 12:20:56|install|texlive-12many-doc|2016.113.0.0.3svn15878-28.1|noarch||repo-oss|fc44bac1a38d05961fdc9d347edcb5dfe0d8c996| 2017-05-10 12:20:56|install|texlive-a2ping-doc|2016.113.svn29725-28.1|noarch||repo-oss|5b9730c197f6f4b44eea629d437ad09eef7f3538| 2017-05-10 12:20:56|install|texlive-a4wide-doc|2016.113.svn20943-28.1|noarch||repo-oss|0bc1d9e1d7e00b847e61da267af7b8664422a563| ... -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 2017-06-10 19:03, Peter Mc Donough wrote:
Anyhow, I got rid of all "textlive", marked it "taboo" in Yast, unticked the checkbox "recommnds" in Yast and added solver.onlyRequires = true in /etc/zypp/zypp.conf.
This should avoid all accidential downloads in future.
But it can also mean that maybe sometime you will hit some functionality that doesn't work. -- Cheers / Saludos, Carlos E. R. (from 42.2 x86_64 "Malachite" at Telcontar)
Am 10.06.2017 um 19:16 schrieb Carlos E. R.:
On 2017-06-10 19:03, Peter Mc Donough wrote:
Anyhow, I got rid of all "textlive", marked it "taboo" in Yast, unticked the checkbox "recommnds" in Yast and added solver.onlyRequires = true in /etc/zypp/zypp.conf.
This should avoid all accidential downloads in future.
But it can also mean that maybe sometime you will hit some functionality that doesn't work.
True, I will have to remember the changes. It is somehing for my little troubles collection in my archive. cu Peter -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 2017-06-10 19:55, Peter Mc Donough wrote:
Am 10.06.2017 um 19:16 schrieb Carlos E. R.:
On 2017-06-10 19:03, Peter Mc Donough wrote:
Anyhow, I got rid of all "textlive", marked it "taboo" in Yast, unticked the checkbox "recommnds" in Yast and added solver.onlyRequires = true in /etc/zypp/zypp.conf.
This should avoid all accidential downloads in future.
But it can also mean that maybe sometime you will hit some functionality that doesn't work.
True, I will have to remember the changes. It is somehing for my little troubles collection in my archive.
My preferred solution is to taboo the specific package(s). However, on a small filesystem no recommends is probably quite the appropriate solution, to minimize size. YaST could ask, when you add certain package, to ask the user about the recommends, accept or refuse. And remember the choice, of course. Would be a nice feature. -- Cheers / Saludos, Carlos E. R. (from 42.2 x86_64 "Malachite" at Telcontar)
participants (22)
-
Andreas Schwab
-
Anton Aylward
-
Bernhard Voelker
-
Carlos E. R.
-
Christian Boltz
-
Felix Miata
-
Freek de Kruijf
-
Jan Engelhardt
-
Karl Cheng
-
Knurpht - Gertjan Lettink
-
Larry Finger
-
Manfred Hollstein
-
Nate Graham
-
Neil Rickert
-
nicholas
-
Patrick Shanahan
-
Peter Mc Donough
-
pit
-
Roman Bysh
-
Simon Lees
-
Werner Flamme
-
Werner LEMBERG