[opensuse] 3 partitions vs. 11 partitions
Yesterday I bought a new laptop and installed 11.4. As it is long ago when I installed OpenSuse last time, I followed the suggestion to create only three Linux partitions: Win7 167 GB (shrinked by the Installation CD) Swap 2 GB / 20 GB /home 260 GB (the remainder) Normally I would have created a much larger number of Linux partitions, for instance: Win7 167 GB (shrinked by the Installation CD, as above) Swap 2 GB / 10 GB /usr 20 GB /opt 10 GB /var 5 GB /tmp 5 GB /boot 0.09 GB /srv 60 GB /data 60 GB /local 60 GB /home 52 GB Does that make sense? Or is it only a waste of space, since the partitions cannot be filled completely? Thanks in advance for your suggestions. Best regards, Wolfgang -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Sat, Apr 9, 2011 at 14:08, Wolfgang Mueller wrote:
Yesterday I bought a new laptop and installed 11.4.
As it is long ago when I installed OpenSuse last time, I followed the suggestion to create only three Linux partitions:
Win7 167 GB (shrinked by the Installation CD) Swap 2 GB / 20 GB /home 260 GB (the remainder)
Normally I would have created a much larger number of Linux partitions, for instance:
Win7 167 GB (shrinked by the Installation CD, as above) Swap 2 GB / 10 GB /usr 20 GB /opt 10 GB /var 5 GB /tmp 5 GB /boot 0.09 GB /srv 60 GB /data 60 GB /local 60 GB /home 52 GB
Does that make sense? Or is it only a waste of space, since the partitions cannot be filled completely?
Thanks in advance for your suggestions.
For a laptop... the simple partitioning is a lot easier and more logical for the kind of use it'll get. There's nothing wrong with splitting up into dozens of partitions if you want, but you're not gaining much on a laptop. I can see using an 11 partition scheme like you describe on a server that uses LVM and RAIDs and you need to be able to protect/manage/shuffle data in the various partitions... or where you can add/remove space out of the LVM drive pool. C. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
C wrote:
For a laptop... the simple partitioning is a lot easier and more logical for the kind of use it'll get. There's nothing wrong with splitting up into dozens of partitions if you want, but you're not gaining much on a laptop. I can see using an 11 partition scheme like you describe on a server that uses LVM and RAIDs and you need to be able to protect/manage/shuffle data in the various partitions... or where you can add/remove space out of the LVM drive pool.
I use LVM and multiple partitions on my ThinkPad. It makes it easier to adjust partition sizes when needed. I also have an ext3 partition which contains the Windows "My Documents" folder, making it easy to exchange documents between operating systems. I use ex2IFS to enable Windows access to the ext3 partition. http://www.fs-driver.org/ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
James Knott wrote:
C wrote:
For a laptop... the simple partitioning is a lot easier and more logical for the kind of use it'll get. There's nothing wrong with splitting up into dozens of partitions if you want, but you're not gaining much on a laptop. I can see using an 11 partition scheme like you describe on a server that uses LVM and RAIDs and you need to be able to protect/manage/shuffle data in the various partitions... or where you can add/remove space out of the LVM drive pool.
I use LVM and multiple partitions on my ThinkPad. It makes it easier to adjust partition sizes when needed. I also have an ext3 partition which contains the Windows "My Documents" folder, making it easy to exchange documents between operating systems. I use ex2IFS to enable Windows access to the ext3 partition.
Forgot to mention. Windows, that shared ext3 and /boot partitions are all on primary partions. The LVM is is a logical partition. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
I use LVM and multiple partitions on my ThinkPad. It makes it easier to adjust partition sizes when needed. I also have an ext3 partition which contains the Windows "My Documents" folder, making it easy to exchange documents between operating systems. I use ex2IFS to enable Windows access to the ext3 partition.
Do you have convenient tools to use with LVM? Last time I had to deal with it (6 months ago), it was so tiresome that I swore I will not ever touch it again. Now, granted, it started with someone else's stupidity. Some person at Dell set up a drive with a 4G /home as primary partition, and a 496G / set up as LVM. Don't ask me what on earth they were thinking. I ended up with the task of remedying this setup. Now, if it was traditional partitioning, this would have meant an hour backing things up "just in case" and then resizing and moving partitions. Turned out, it was easy to increase a volume size, but painfully difficult to decrease it. GParted wouldn't do it, and neither would any tools included with RHEL. I had limited internet access, but anyway Google search did not produce any obvious tools that could do the job. I ended up having to do manual calculations on cylinder level for all the resizing, and hope to God that I didn't make any errors in my arithmetic. My general conclusion was, yes, I was fixing something that was completely mad to start with. But this madness would have been easier to fix in the bad old traditional system, simply because it has better tools to go with it. So I am not touching LVM until I know of easy tools to use with it that would deal with a wide range of possible configurations, even crazy ones ;-) Myrosia -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Myrosia Dzikovska wrote:
Do you have convenient tools to use with LVM? Last time I had to deal with it (6 months ago), it was so tiresome that I swore I will not ever touch it again.
No, just what's in Yast. I create partitions somewhat bigger than currently needed and add to them as necessary. However, I do find creating LVM partitions to be a bit awkward. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
James Knott said the following on 04/09/2011 04:23 PM:
Myrosia Dzikovska wrote:
Do you have convenient tools to use with LVM? Last time I had to deal with it (6 months ago), it was so tiresome that I swore I will not ever touch it again.
No, just what's in Yast. I create partitions somewhat bigger than currently needed and add to them as necessary. However, I do find creating LVM partitions to be a bit awkward.
There's a tool in YAST? No in mine. # lvcreate -L 1G -n downloads vgmain # mkfs.reiserfs /dev/vgmain/downloads Really awkward, eh? Not big enough? # lvextend -L+1G /dev/vgmain/downloads # resize_reiserfs -s+1G -f /dev/vgmain/downloads Oops, too much # umount /dev/vgmain/downloads # resize_resiserfs -s-512M /dev/vgmain/downloads # lvreduce -L -512M /dev/vgmain/downloads # mount /dev/vgmain/downloads Of course not all file systems can be extended and reduced like that. You may have to take a backup, easy to do with the LVM snapshot mechanism, then delete the original, create new and copy back. But wouldn't it be easier to use a better file system? -- People who make no mistakes do not usually make anything. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Anton Aylward wrote:
There's a tool in YAST? No in mine.
On my 11.0 system, in Yast, there's "LVM" under System. On one computer I have here (IBM Netfinity server) I have 4 SCSI drives, so I first create a RAID array and place the LVM on it. The only computers I don't use LVM on are my firewall and a test system, where I'm not too worried about disk space. Also, my main system (the one I'm using right now) has one drive containing only /home, so no LVM for it. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Myrosia Dzikovska said the following on 04/09/2011 04:09 PM:
Do you have convenient tools to use with LVM? Last time I had to deal with it (6 months ago), it was so tiresome that I swore I will not ever touch it again.
Now, granted, it started with someone else's stupidity. Some person at Dell set up a drive with a 4G /home as primary partition, and a 496G / set up as LVM. Don't ask me what on earth they were thinking. I ended up with the task of remedying this setup.
Was you problem with resizing the partitions or resizing the file systems on the partitions? Trying to shrink that 496G root partition without first shrinking the file system on it would lead to disaster! The tool I used with IBM AIX did both "under the hood". The paper I referred to by Michael Hasenstein explains quite clearly how to both shrink and grow partitions with file systems on them. Do realise that some file systems cannot be resized. Michael also describes a YAST1 gui for doing LVM management. See also http://www.linuxtopia.org/online_books/opensuse_guides/opensuse11.1_referenc... You might also look at /sbin/fsadm Which is incorrect: you can resize a ReiserFS without unmounting. I've done it many times :-) -- "What we have learned from others becomes our own by reflection". -- Ralph Waldo Emerson -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Saturday 09 of April 2011 15:08:00 Wolfgang Mueller wrote:
As it is long ago when I installed OpenSuse last time, I followed the suggestion to create only three Linux partitions:
Win7 167 GB (shrinked by the Installation CD) Swap 2 GB / 20 GB /home 260 GB (the remainder)
My configuration is: swap 2GB / 60 GB /home 233 GB Both partitions are on LVM. I see no reason to change it.
Thanks in advance for your suggestions.
Best regards, Wolfgang
Regards, Peter -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
2011/4/9
On Saturday 09 of April 2011 15:08:00 Wolfgang Mueller wrote:
As it is long ago when I installed OpenSuse last time, I followed the suggestion to create only three Linux partitions:
Win7 167 GB (shrinked by the Installation CD) Swap 2 GB / 20 GB /home 260 GB (the remainder)
My configuration is: swap 2GB / 60 GB /home 233 GB
Both partitions are on LVM. I see no reason to change it.
Thanks in advance for your suggestions.
Best regards, Wolfgang
Regards, Peter
Soooo, we are not talking about partitions.... This days is nonsense to use plain old partitions... Using LVM I create filesystems with only the needed space (create as many as you see fit), and as needed extend the volumes and FS. Making big fat FS without even knowing if you're going to fill it defeats the purpose of using LVM, always leave unassigned space to grow current volumes or create new ones... Regards, -- Ciro Iriarte http://cyruspy.wordpress.com -- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On 9 April 2011 20:39,
On Saturday 09 of April 2011 15:08:00 Wolfgang Mueller wrote:
As it is long ago when I installed OpenSuse last time, I followed the suggestion to create only three Linux partitions:
Win7 167 GB (shrinked by the Installation CD) Swap 2 GB / 20 GB /home 260 GB (the remainder)
My configuration is: swap 2GB / 60 GB /home 233 GB
for a laptop ,auxsvr's partion is a good one (makes me happy ^_^), but / is a little big ,I think . 50G(even less then 50) is quite enough , we(maybe I) don't need so many space actually ~
Both partitions are on LVM. I see no reason to change it.
Thanks in advance for your suggestions.
Best regards, Wolfgang
Regards, Peter -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-- ________________________ wolf python london(WPL) Do as you soul should do ! ________________________ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On 04/09/2011 08:08 AM, Wolfgang Mueller pecked at the keyboard and wrote:
Yesterday I bought a new laptop and installed 11.4.
As it is long ago when I installed OpenSuse last time, I followed the suggestion to create only three Linux partitions:
Win7 167 GB (shrinked by the Installation CD) Swap 2 GB / 20 GB /home 260 GB (the remainder)
Normally I would have created a much larger number of Linux partitions, for instance:
Win7 167 GB (shrinked by the Installation CD, as above) Swap 2 GB / 10 GB /usr 20 GB /opt 10 GB /var 5 GB /tmp 5 GB /boot 0.09 GB /srv 60 GB /data 60 GB /local 60 GB /home 52 GB
Does that make sense? Or is it only a waste of space, since the partitions cannot be filled completely?
Partition schemes are like belly buttons, everyone has one and no two are the same. Use whatever makes you happy. What makes *me* happy: swap: 2 GB /boot 200 MB / the rest of the disk -- Ken Schneider SuSe since Version 5.2, June 1998 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Ken Schneider - openSUSE said the following on 04/09/2011 08:55 AM:
What makes *me* happy:
swap: 2 GB /boot 200 MB / the rest of the disk
Makes my *unhappy* I really like having a /tmp so that files there can be * noexec, nosetuid * sticky for the user that created them * avoid hard links to /sbin and /usr/sbin and some vulnerabilities that go with that. I really like having a /home, /var and /tmp so that a runaway user application can't consume all the space on / and /root so that I can log in as root and kill it off. YMMV. I'm a bit paranoid about these things. -- The whole problem with the world is that fools and fanatics are always so certain of themselves, and wiser people so full of doubts. --Bertrand Russell -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Wolfgang Mueller said the following on 04/09/2011 08:08 AM:
Does that make sense? Or is it only a waste of space, since the partitions cannot be filled completely?
No. First, I would shrink windows more, but then I have only one windows application I ever use. You may be different, so that's just an observation. Second, because we are stuck with grub-0.97 you need a real (as in primary partition) /boot. If ThePowersThatBe would get their act together and move to grub-2 we could all do this properly. By properly I mean setting ALL of your drive to run under LVM. As it is you can;'t do that. You need a 'real" /boot and /. They don't have to be that big. I have a 150M /boot on /dev/sda4 and that has two generations of kernels and is less than 50% used. I have a 2G / on /dev/sda6 that is only about 20% used. I have a small, nominal swap partition of 56000 blocks. I have /dev/shm /var/lock and /var/run as tmpfs I'm thinking of putting /tmp as a tmpfs, but on as laptop I am more memory limited than on a desktop motherboard. Everything else on my laptop is managed under LVM2. For the most part I run ReiserFS since its easier to grow and shrink than ext3. The thing is that I seem to have more smaller partitions. Many seem to work out at or add up to around 4G, so I can back them up onto a DVD :-) For example, I have /home, and when that filled up I created /home/anton/Documents /dev/mapper/vgmain-Documents 600M 36% /home/anton/.thunderbird /dev/mapper/vgmain-Thunderbird 1.7G 35% and so on. I have /usr, /usr/share, /usr/lib, /usr/lib/ruby, /usr/src and /home/anton/Media as the most notable. The other root level directories you mention are there but work out small. I have nothing over 75% fill - /usr/share Most partitions are under 50%. As a result I rarely need to tweak things now, buy a couple of years ago I had to create .thunderbird as /home filled up and /usr/lib/ruby and /home/anton/Ruby as I started working with Ruby. Since I have a minimalist approach - I only allocate what's needed plus a margin - my Linux installation works out at about 60G. So I have a LOT of spare capacity. I use some of that for taking snapshots, something that make backups a joy and something that's impossible with tradition partition methods. To be fair, a lot of what might take up space or other people is on my file server and is NFS mounted. Buy my point is that if you use LVM you are not "fixed" when you partition your drive. If I want to set up a major web application and need to grow /srv that is just two commands: one to grow the partition, one to grow the file system. Similarly if I need to grow the /data where the MySQL database lives. What annoys me is that openSuse does not have a good LVM management gui. SuSe has had LVM since 6.3 and Michael Hasenstein wrote a great paper back in 2001. http://www.suse.de/cgi-bin/print_page_www.pl?NPSPath=/webredesign/htdocs/en/... See also http://stommel.tamu.edu/~baum/programming.html#LVM In this day and age, a fixed partition installation for anything except an embedded and/or limited system simply doesn't make sense. -- A hollow voice says, "A hollow voice says, 'Plugh'." -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On 2011/04/09 14:08 (GMT+0200) Wolfgang Mueller composed:
Yesterday I bought a new laptop and installed 11.4.
As it is long ago when I installed OpenSuse last time, I followed the suggestion to create only three Linux partitions:
Win7 167 GB (shrinked by the Installation CD)
WinVista & Win7 have their own resizer built in. Many find fewer opportunities for unwanted surprises using it instead of non-native resizers.
Swap 2 GB / 20 GB /home 260 GB (the remainder)
I guess you always know ahead of time that at upgrade time the new version will be satisfactory, so you can blow away the old one without trying the new one first, or using the old one to repair a mistake in the new one, or vice versa. Not me. I almost never partition for less than two / partitions except on a very small HD system. It's way easier for me to do repairs from another installed system than hunt around for or run from a live CD or DVD. If I don't like the "upgrade", I simply don't use it, no reversion necessary. Note too that on most HDs, the fastest part of the disk is the front, so you're giving Win7 a speed advantage leaving all of it at the front. You could start all over, giving Win7 a small boot partition at the front, and installing its OS someplace slower, and/or dividing it into separate OS and home partitions, so as not to have to lose your win data when Windows needs re-installing. cf. http://fm.no-ip.com/PC/install-doz-after.html -- "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+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Sat, 2011-04-09 at 14:08 +0200, Wolfgang Mueller wrote:
Yesterday I bought a new laptop and installed 11.4.
As it is long ago when I installed OpenSuse last time, I followed the suggestion to create only three Linux partitions:
Win7 167 GB (shrinked by the Installation CD) Swap 2 GB / 20 GB /home 260 GB (the remainder)
Normally I would have created a much larger number of Linux partitions, for instance:
Win7 167 GB (shrinked by the Installation CD, as above) Swap 2 GB / 10 GB /usr 20 GB /opt 10 GB /var 5 GB /tmp 5 GB /boot 0.09 GB /srv 60 GB /data 60 GB /local 60 GB /home 52 GB
Does that make sense? Or is it only a waste of space, since the partitions cannot be filled completely?
Thanks in advance for your suggestions.
Like some said before, put all (except /boot) in logical volumes. Normally i keep /usr and /opt at about 80% usage. I used to make /boot 100MB, biut it seems that its not enough, so double it to 200MB Considering srv,data and local: data belonging to server processes (apache, mysql, openldap, ftp, tftp and so on should reside under /srv) Seems rather large for a laptop imho, so i should shrink it to what you need. It is the nice thing about lvm, you can enlarge them when you need it, thus not wasting space, but atoh keep things seperate that could interfere In case you run processes that "complain" a lot, you might consider putting /var/log into a separate partition. Some even put /var/run into a tiny seperate partition (some MB) so that they have a garanteed working area to write their PID Oh, and btw, if you take the laptop along, don't forget to add encryption to the volume-group. hw -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Hans Witvliet said the following on 04/09/2011 09:52 AM:
In case you run processes that "complain" a lot, you might consider putting /var/log into a separate partition. Some even put /var/run into a tiny seperate partition (some MB) so that they have a garanteed working area to write their PID
Its easy enough (go google) to put /var/run and /var/lock in a tmpfs. Its not that difficult to put /tmp in a tmpfs if you think you can spare the ram/swap. The tmpfs file system is even more of "an elastic band" than LVM! And the nice thing is that it gets purged on reboot! There is a slight speed advantage in not hitting the disk for /var/run and /var/lock, but you'd be hard pressed to see it, not least of all in that if you're that hard pressed you're probably consuming ram and pushing it out onto swap and so getting back the disk activity. But having /tmp as a tmpfs *does* make a difference Years ago, Mike Tilson of HCR wrote a driver for SCO XENIX (if you can remember that far back) which mapped the i-nodes of the V7fs into the high ram on a PDP-11 that the OS itself couldn't make use of :-) Boy! did that make a difference! Later model file systems aren't quite that easy to map and later model Linux makes better use of address space (and if it doesn't, then use -pae), so we end up with tmpfs. Its a winner.And its easy to use. There are fixes out there for mapping /var/log to a tmpfs, but that makes me a bit nervous; I like to be able to retain logs across boots. -- Let us never forget that government is ourselves and not an alien power over us. The ultimate rulers of our democracy are not a President and senators and congressmen and government officials, but the voters of this country.” -- Franklin D. Roosevelt -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Sun, 10 Apr 2011 01:09:39 Anton Aylward wrote:
[...] There is a slight speed advantage in not hitting the disk for /var/run and /var/lock, but you'd be hard pressed to see it, not least of all in that if you're that hard pressed you're probably consuming ram and pushing it out onto swap and so getting back the disk activity.
But having /tmp as a tmpfs *does* make a difference [...]
Thanks for the tip on this one - I just changed mine (/tmp was on /) and, man, did it make a difference to how kde 4 runs?!! Almost everything is noticeably snappier (especially Dolphin!). Firefox seems better too.
[...] There are fixes out there for mapping /var/log to a tmpfs, but that makes me a bit nervous; I like to be able to retain logs across boots.
Easy - mount -t tmpfs -o size=100MB /var/log I hear your comment about retaining logs across boots, though. Sometimes that's needed, but the uptime on my system currently sits at 93 days, 10:50 (and I can't actually remember what that reboot was for...I think it may have been a power outage, actually)... Cheers, Rodney. -- =================================================== Rodney Baker VK5ZTV rodney.baker@iinet.net.au =================================================== -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On 2011/04/10 03:34 (GMT+0930) Rodney Baker composed:
Anton Aylward wrote:
But having /tmp as a tmpfs *does* make a difference
Thanks for the tip on this one - I just changed mine (/tmp was on /) and, man, did it make a difference to how kde 4 runs?!!
My brain isn't connecting with what to actually do in fstab to accomplish this. Right now I have: LABEL=07tmp /tmp /ext2 noatime,noacl 1 2 What would I do instead?
Easy - mount -t tmpfs -o size=100MB /var/log -- "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+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Sun, 10 Apr 2011 03:20:15 Felix Miata wrote:
On 2011/04/10 03:34 (GMT+0930) Rodney Baker composed:
Anton Aylward wrote:
But having /tmp as a tmpfs *does* make a difference
Thanks for the tip on this one - I just changed mine (/tmp was on /) and, man, did it make a difference to how kde 4 runs?!!
My brain isn't connecting with what to actually do in fstab to accomplish this. Right now I have:
LABEL=07tmp /tmp /ext2 noatime,noacl 1 2
What would I do instead?
Try it out first with: mount -t tmpfs /tmp [-o size=<value>] Then if you want to make it permanent add the following to /etc/fstab: none /tmp tmpfs size=<VALUE> 0 0 Size is optional - it defaults to 50% of RAM. It accepts either a value in bytes (e.g. 200m for 200MB) or a % (of RAM) - mine is set to 25% as a starting point. BTW, size is the *limit* - it will not grow past that (but it will only use what it needs otherwise). -- =================================================== Rodney Baker VK5ZTV rodney.baker@iinet.net.au =================================================== -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Saturday, April 09, 2011 08:40:47 pm Rodney Baker wrote:
Try it out first with:
mount -t tmpfs /tmp [-o size=<value>]
I needed to use mount -t tmpfs [-o size=<value>] tmpfs /tmp -- Don -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On 4/9/2011 10:40 PM, Rodney Baker wrote:
On Sun, 10 Apr 2011 03:20:15 Felix Miata wrote:
On 2011/04/10 03:34 (GMT+0930) Rodney Baker composed:
Anton Aylward wrote:
But having /tmp as a tmpfs *does* make a difference
Thanks for the tip on this one - I just changed mine (/tmp was on /) and, man, did it make a difference to how kde 4 runs?!!
My brain isn't connecting with what to actually do in fstab to accomplish this. Right now I have:
LABEL=07tmp /tmp /ext2 noatime,noacl 1 2
What would I do instead?
Try it out first with:
mount -t tmpfs /tmp [-o size=<value>]
Then if you want to make it permanent add the following to /etc/fstab:
none /tmp tmpfs size=<VALUE> 0 0
Size is optional - it defaults to 50% of RAM. It accepts either a value in bytes (e.g. 200m for 200MB) or a % (of RAM) - mine is set to 25% as a starting point. BTW, size is the *limit* - it will not grow past that (but it will only use what it needs otherwise).
This should be good... why, pray tell, would you put swap in ram? -- bkw -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Brian K. White wrote:
This should be good... why, pray tell, would you put swap in ram?
Yes, that argument ought to be really entertaining to see. -- Per Jessen, Zürich (19.8°C) -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On 4/9/2011 1:34 PM, Rodney Baker wrote:
On Sun, 10 Apr 2011 01:09:39 Anton Aylward wrote:
[...] There is a slight speed advantage in not hitting the disk for /var/run and /var/lock, but you'd be hard pressed to see it, not least of all in that if you're that hard pressed you're probably consuming ram and pushing it out onto swap and so getting back the disk activity.
But having /tmp as a tmpfs *does* make a difference [...]
Thanks for the tip on this one - I just changed mine (/tmp was on /) and, man, did it make a difference to how kde 4 runs?!!
Almost everything is noticeably snappier (especially Dolphin!). Firefox seems better too.
[...] There are fixes out there for mapping /var/log to a tmpfs, but that makes me a bit nervous; I like to be able to retain logs across boots.
Easy - mount -t tmpfs -o size=100MB /var/log
I hear your comment about retaining logs across boots, though. Sometimes that's needed, but the uptime on my system currently sits at 93 days, 10:50 (and I can't actually remember what that reboot was for...I think it may have been a power outage, actually)...
Of course you don't need to fix your roof when it's not raining. The day your box DOES reboot (unexpectedly) is one of the days you want the logs leading up to the event THE MOST. -- bkw -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Brian K. White said the following on 04/09/2011 07:26 PM:
The day your box DOES reboot (unexpectedly) is one of the days you want the logs leading up to the event THE MOST.
+1 -- Faith may be defined briefly as an illogical belief in the occurrence of the improbable...A man full of faith is simply one who has lost (or never had) the capacity for clear and realistic thought. He is not a mere ass: he is actually ill. -- H. L. Mencken -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Sun, 10 Apr 2011 08:56:41 Brian K. White wrote:
On 4/9/2011 1:34 PM, Rodney Baker wrote:
On Sun, 10 Apr 2011 01:09:39 Anton Aylward wrote:
[...] There is a slight speed advantage in not hitting the disk for /var/run and /var/lock, but you'd be hard pressed to see it, not least of all in that if you're that hard pressed you're probably consuming ram and pushing it out onto swap and so getting back the disk activity.
But having /tmp as a tmpfs *does* make a difference [...]
Thanks for the tip on this one - I just changed mine (/tmp was on /) and, man, did it make a difference to how kde 4 runs?!!
Almost everything is noticeably snappier (especially Dolphin!). Firefox seems better too.
[...] There are fixes out there for mapping /var/log to a tmpfs, but that makes me a bit nervous; I like to be able to retain logs across boots.
Easy - mount -t tmpfs -o size=100MB /var/log
I hear your comment about retaining logs across boots, though. Sometimes that's needed, but the uptime on my system currently sits at 93 days, 10:50 (and I can't actually remember what that reboot was for...I think it may have been a power outage, actually)...
Of course you don't need to fix your roof when it's not raining.
The day your box DOES reboot (unexpectedly) is one of the days you want the logs leading up to the event THE MOST.
Absolutely, which is why I *don't* have /var/log on tmpfs. :-) -- =================================================== Rodney Baker VK5ZTV rodney.baker@iinet.net.au =================================================== -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On 4/9/2011 1:34 PM, Rodney Baker wrote:
On Sun, 10 Apr 2011 01:09:39 Anton Aylward wrote:
[...] There is a slight speed advantage in not hitting the disk for /var/run and /var/lock, but you'd be hard pressed to see it, not least of all in that if you're that hard pressed you're probably consuming ram and pushing it out onto swap and so getting back the disk activity.
But having /tmp as a tmpfs *does* make a difference [...]
Thanks for the tip on this one - I just changed mine (/tmp was on /) and, man, did it make a difference to how kde 4 runs?!!
Almost everything is noticeably snappier (especially Dolphin!). Firefox seems better too.
[...] There are fixes out there for mapping /var/log to a tmpfs, but that makes me a bit nervous; I like to be able to retain logs across boots.
Easy - mount -t tmpfs -o size=100MB /var/log
I hear your comment about retaining logs across boots, though. Sometimes that's needed, but the uptime on my system currently sits at 93 days, 10:50 (and I can't actually remember what that reboot was for...I think it may have been a power outage, actually)...
Of course you don't need to fix your roof when it's not raining. The day your box DOES reboot (unexpectedly) is one of the days you want the logs leading up to the event THE MOST. -- bkw -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On 04/09/2011 07:08 AM, Wolfgang Mueller wrote:
Yesterday I bought a new laptop and installed 11.4.
As it is long ago when I installed OpenSuse last time, I followed the suggestion to create only three Linux partitions:
Win7 167 GB (shrinked by the Installation CD) Swap 2 GB / 20 GB /home 260 GB (the remainder)
Normally I would have created a much larger number of Linux partitions, for instance:
Win7 167 GB (shrinked by the Installation CD, as above) Swap 2 GB / 10 GB /usr 20 GB /opt 10 GB /var 5 GB /tmp 5 GB /boot 0.09 GB /srv 60 GB /data 60 GB /local 60 GB /home 52 GB
Does that make sense? Or is it only a waste of space, since the partitions cannot be filled completely?
Thanks in advance for your suggestions.
Best regards, Wolfgang
I have done it both ways (lots of partitions) and (normal 3 /boot, /, /home). Eg: 15:29 archangel:~> cat /proc/partitions major minor #blocks name 8 16 732574584 sdb 8 32 488386584 sdc 8 48 732574584 sdd 8 80 1465138584 sdf 8 81 1 sdf1 8 85 1465136128 sdf5 8 64 1465138584 sde 8 65 1 sde1 8 69 1465136128 sde5 8 0 488386584 sda 254 0 488386583 dm-0 254 1 732574583 dm-1 254 2 72229 dm-2 254 3 2104483 dm-3 254 4 20972826 dm-4 254 5 465234336 dm-5 254 6 19534977 dm-6 254 7 120456 dm-7 254 8 39062016 dm-8 254 9 1951866 dm-9 254 10 15358108 dm-10 254 11 30716248 dm-11 254 12 7510356 dm-12 254 13 618317721 dm-13 ...and... 15:30 supersff:~/tblds/qt3-45> cat /proc/partitions major minor #blocks name 8 0 488386584 sda 8 1 71681998 sda1 8 2 1 sda2 8 5 286720056 sda5 8 6 30716248 sda6 8 7 1951866 sda7 8 8 1261071 sda8 Sooner or later you will learn that the more partitions you have -- the more you will screw yourself due to the inability to create softlinks across partition boundaries. Don't get me wrong, I see great advantages from a backup/imaging standpoint of multiple partitions. But, I can backup regardless of how the partitions are laid out -- but I can't softlink across a partition boundary no matter how bad I "really really want to." KISS - is the best philosophy. Use only the number that you need. If you need 50 - use 50, if you can really get by with the normal 3 -- then you are better off with the normal 3. -- David C. Rankin, J.D.,P.E. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On 04/09/2011 01:33 PM, David C. Rankin wrote:
Sooner or later you will learn that the more partitions you have -- the more you will screw yourself due to the inability to create softlinks across partition boundaries. Don't get me wrong, I see great advantages from a backup/imaging standpoint of multiple partitions. But, I can backup regardless of how the partitions are laid out -- but I can't softlink across a partition boundary no matter how bad I "really really want to."
KISS - is the best philosophy. Use only the number that you need. If you need 50 - use 50, if you can really get by with the normal 3 -- then you are better off with the normal 3.
I have to agree with this. I've gravitated to 3 (root, swap, home), for personal machines and 4 for production servers (adding DATA partition for shared mission critical data). I'm well aware of the potential for some process filling the root partition with runaway logging (or some such) but quite frankly that hasn't happened to me since the Pleistocene. -- Explain again the part about rm -rf / -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Sat, 2011-04-09 at 13:44 -0700, jsa wrote:
On 04/09/2011 01:33 PM, David C. Rankin wrote:
Sooner or later you will learn that the more partitions you have -- the more you will screw yourself due to the inability to create softlinks across partition boundaries. Don't get me wrong, I see great advantages from a backup/imaging standpoint of multiple partitions. But, I can backup regardless of how the partitions are laid out -- but I can't softlink across a partition boundary no matter how bad I "really really want to."
KISS - is the best philosophy. Use only the number that you need. If you need 50 - use 50, if you can really get by with the normal 3 -- then you are better off with the normal 3.
I have to agree with this. I've gravitated to 3 (root, swap, home), for personal machines and 4 for production servers (adding DATA partition for shared mission critical data).
I'm well aware of the potential for some process filling the root partition with runaway logging (or some such) but quite frankly that hasn't happened to me since the Pleistocene.
So apperantly you never raised log-levels for openldap, openvpn etc. These can produce _huge_ amount of log-info I know that apache is quite modest with log-info, but if you for get to activate log-rotate (or in my case: let others create theit own virtual apache-servers + with their own logs) it still can fill up places you don't want to be filled up. hw -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
David C. Rankin wrote:
but I can't softlink across a partition boundary no matter how bad I "really really want to."
I thought it was hard links that didn't work between partitions. Soft links certainly do. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Sat, 2011-04-09 at 15:33 -0500, David C. Rankin wrote:
Sooner or later you will learn that the more partitions you have -- the more you will screw yourself due to the inability to create softlinks across partition boundaries. Don't get me wrong,
With _soft_ links you do can span across different file systems. Only with hard links you can not. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
David C. Rankin said the following on 04/09/2011 04:33 PM:
I can backup regardless of how the partitions are laid out -- but I can't softlink across a partition boundary no matter how bad I "really really want to."
Are you sure you are saying what you mean? I understand not being able to HARD link across a partition boundary, but I have a *LOT*of soft links. In fact I thought the whole point of soft links was that they could go across a partition boundary. For example: # ls -l /etc/termcap lrwxrwxrwx 1 root root 23 2010-07-21 22:12 /etc/termcap -> /usr/share/misc/termcap # df /etc /usr/share/misc Filesystem 1K-blocks Used Available Use% Mounted on /dev/sda6 2205780 568688 1525044 28% / /dev/mapper/vgmain-usrshare 3145628 2354040 791588 75% /usr/share See? Actually I have symlinks from my desktop to many places on my various mounted file systems. Some users might recall having used 'thin clients' or 'roaming login' where the home directory is actually a symlink to /mnt/nft/server/<username> which is in turn mounted by the automounter or a PAM module such as "pam_csync" or "pam_mount".
From man ln(1)
... Create hard links by default, symbolic links with --symbolic. ... See also symlink(2), synlink(7), readlink, ln(1p) -- Ah, nostalgia ain't what it used to be. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
David C. Rankin said the following on 04/09/2011 04:33 PM:
KISS - is the best philosophy. Use only the number that you need.
And that's the wonderful thing about LVM. If I need to compile a package I can create /usr/src and give that space back when I'm finished then next week use that same space as /data to do some analytics ...or to take a snapshot of other partitions in turn for backup. See how simple this last function can be at http://tldp.org/HOWTO/LVM-HOWTO/snapshots_backup.html More detail at http://www.howtoforge.com/linux_lvm_snapshots <quote> An LVM snapshot is an exact copy of an LVM partition that has all the data from the LVM volume from the time the snapshot was created. The big advantage of LVM snapshots is that they can be used to greatly reduce the amount of time that your services/databases are down during backups because a snapshot is usually created in fractions of a second. After the snapshot has been created, you can back up the snapshot while your services and databases are in normal operation. </quote> Applied to backing up a MySQL database http://blog.dbadojo.com/2007/09/mysql-backups-using-lvm-snapshots.html Of course you can also create LVM groups dynamcially with loopback http://www.howtoforge.com/creating-portable-disksafes-with-loopbackfs-and-lv... -- Wisdom is earned through bitter experience. Idiocy comes easily. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Hello, On Sun, 10 Apr 2011, Anton Aylward wrote:
David C. Rankin said the following on 04/09/2011 04:33 PM:
KISS - is the best philosophy. Use only the number that you need.
And that's the wonderful thing about LVM.
KISS is exactly why I don't use LVM. $ grep 'sd.[0-9]$' /proc/partitions 8 17 31457280 sdb1 8 18 1048576 sdb2 8 19 31457280 sdb3 8 20 1 sdb4 8 21 1889550304 sdb5 8 1 5253223 sda1 8 2 1060290 sda2 8 3 31463302 sda3 8 4 1 sda4 8 5 5253223 sda5 8 6 31463271 sda6 8 7 413890596 sda7 8 49 1953514552 sdd1 8 65 1953513560 sde1 8 81 1953514552 sdf1 8 97 1044188 sdg1 8 98 10490448 sdg2 8 99 1941977362 sdg3 8 113 1953514552 sdh1 8 33 1465136001 sdc1 8 129 488384001 sdi1 How do you sanely handle this? When disks die? Oh, and there's also another 6 disks (or more) imported via NFS when needed, and 3 "good" external (and a couple "scrap"[3] ones). I mostly use a tree of symlinks (including the NFS-mounts when available) basically starting in ~. Let's say I'd use LVM, what if e.g. /dev/sde shows defects (seen in the SMART data). Can LVM show me what dirs and files(!) I have on the PV(s) on that probably soon-failing drive?
If I need to compile a package I can create /usr/src and give that space back when I'm finished then next week use that same space as /data to do some analytics ...or to take a snapshot of other partitions in turn for backup.
For stuff like that, I've got a bunch of GB of scrap-space (e.g. in /data/~user or /home2/~user/[4]) where I can create a dir, if needed symlink or 'mount --bind' it to anywhere I want, use it and then remove the whole stuff again. For small stuff, /dev/shm/~user suffices. No need to add a level of complex software (AHS, ASS![1] and Murphy![2]) that can go wrong. -dnh, hand-selected, translated sig for this occasion [1] All Hardware Sucks, All Software Sucks, [2] anything that can go wrong, will. At the most inconvenient time. Wreaking as much havoc as possible [e.g. taking the most important data with it]. [3] i.e. ones that have a few defective sectors, but still work and don't rapidly get more defects [4] containing big, but less important stuff, think thumbnails, savegames giggle-mudball cache and suchlike. -- RAID: One more disk fails than can be recovered by the redundancy. -- Andreas Dau -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
David Haller wrote:
Let's say I'd use LVM, what if e.g. /dev/sde shows defects (seen in the SMART data). Can LVM show me what dirs and files(!) I have on the PV(s) on that probably soon-failing drive?
You can instruct LVM to stop using /dev/sde and to start migrating data away. -- Per Jessen, Zürich (9.6°C) -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
David Haller wrote:
Let's say I'd use LVM, what if e.g. /dev/sde shows defects (seen in the SMART data). Can LVM show me what dirs and files(!) I have on the PV(s) on that probably soon-failing drive?
Doesn't matter because in the sane world you're now running LVM over a RAID, right? So the RAID tells you it's replaced /dev/sde with the hot spare and the rebuild is in progress. End of story. :-P
-dnh, hand-selected, translated sig for this occasion
[1] All Hardware Sucks, All Software Sucks,
[2] anything that can go wrong, will. At the most inconvenient time. Wreaking as much havoc as possible [e.g. taking the most important data with it].
But yes, I do agree with these points. Sometimes it would keep me awake at night if I was the obsessive worrying type. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Dave Howorth said the following on 04/12/2011 09:37 AM:
[1] All Hardware Sucks, All Software Sucks,
[2] anything that can go wrong, will. At the most inconvenient time. Wreaking as much havoc as possible [e.g. taking the most important data with it]. But yes, I do agree with these points. Sometimes it would keep me awake at night if I was the obsessive worrying type.
Indeed. But that's why we engineer (aka "design and build") for resilience and to be able to manage failures. -- Fear not those who argue, but those who dodge. -- Marie Ebner von Eschenbach -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Hello, On Tue, 12 Apr 2011, Dave Howorth wrote:
David Haller wrote:
Let's say I'd use LVM, what if e.g. /dev/sde shows defects (seen in the SMART data). Can LVM show me what dirs and files(!) I have on the PV(s) on that probably soon-failing drive?
Doesn't matter because in the sane world you're now running LVM over a RAID, right? So the RAID tells you it's replaced /dev/sde with the hot spare and the rebuild is in progress. End of story. :-P
Ok, now, how'd you contruct a raid over my discs? And note that so many having 2T size is purely coincidental, due to the fact that available and sanely prices sizes (compared to 2T) have stalled there for over a year! 18 months ago, I had about 2 of each of 500G, 750G, 1000G, 1500G, 2000G or so. Now, how do you make a sane (Linux Software-)RAID (Levels 5 or 6), over _those_ drives? And note: the steady state of (my) disks is full! It's been YEARS since I've been below ~85% usage overall (and that only with a new, fresh and bigger HDD installed). Yes, those 5%-15% "free" amount to TB currently, but the problem remains. The simple fact is, I can't shell out to replace e.g. 6 500G drives by 6 1T drives, then 6 2T drives etc. I HAVE differently sized disks. How to have a sane RAID thingy across those without going bonkers? I'd be real happy if you'd come up with a good idea!
-dnh, hand-selected, translated sig for this occasion
[1] All Hardware Sucks, All Software Sucks,
[2] anything that can go wrong, will. At the most inconvenient time. Wreaking as much havoc as possible [e.g. taking the most important data with it].
But yes, I do agree with these points. Sometimes it would keep me awake at night if I was the obsessive worrying type.
If I were, I'd have twice as much disks, each one bought with a size-twin from a different vendor, each possibly mirrored to it's twin ... Or something ... -dnh, who's lost 2 500G drives a while ago, one announced itself, I got most copied off of it, one dying "just like that" (not even showed up in the BIOS anymore, plain dead). Anyway, "re-gettable" data. Still missing some stuff, but it's just TV in that case ;) Irregettable stuff is kept on at least two drives, occasionally copied to the other box to yet a third drive. I'd be grilled though if there'd be a flood. -- [Die Merkel] sitzt dann in der VIP-Loge beim Fußball rum in ihrem rosa Pokemon Kostüm ... -- Hagen Rether -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
David Haller wrote:
Hello,
On Tue, 12 Apr 2011, Dave Howorth wrote:
David Haller wrote:
Let's say I'd use LVM, what if e.g. /dev/sde shows defects (seen in the SMART data). Can LVM show me what dirs and files(!) I have on the PV(s) on that probably soon-failing drive?
Doesn't matter because in the sane world you're now running LVM over a RAID, right? So the RAID tells you it's replaced /dev/sde with the hot spare and the rebuild is in progress. End of story. :-P
Ok, now, how'd you contruct a raid over my discs? And note that so many having 2T size is purely coincidental, due to the fact that available and sanely prices sizes (compared to 2T) have stalled there for over a year! 18 months ago, I had about 2 of each of 500G, 750G, 1000G, 1500G, 2000G or so. Now, how do you make a sane (Linux Software-)RAID (Levels 5 or 6), over _those_ drives?
I would probably make a number of RAID1 arrays of equal-size disks, then use LVM for managing the space. If I needed RAID6, I would have to do with the smallest disk-size, but otherwise I don't see a problem. -- Per Jessen, Zürich (9.8°C) -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Hello, sorry it took me so long to reply. On Tue, 12 Apr 2011, Per Jessen wrote:
David Haller wrote:
On Tue, 12 Apr 2011, Dave Howorth wrote:
David Haller wrote:
Let's say I'd use LVM, what if e.g. /dev/sde shows defects (seen in the SMART data). Can LVM show me what dirs and files(!) I have on the PV(s) on that probably soon-failing drive?
Doesn't matter because in the sane world you're now running LVM over a RAID, right? So the RAID tells you it's replaced /dev/sde with the hot spare and the rebuild is in progress. End of story. :-P
Ok, now, how'd you contruct a raid over my discs? And note that so many having 2T size is purely coincidental, due to the fact that available and sanely prices sizes (compared to 2T) have stalled there for over a year! 18 months ago, I had about 2 of each of 500G, 750G, 1000G, 1500G, 2000G or so. Now, how do you make a sane (Linux Software-)RAID (Levels 5 or 6), over _those_ drives?
I would probably make a number of RAID1 arrays of equal-size disks, then use LVM for managing the space.
RAID1 as underlying RAID-Level would be to wasteful for my data. RAID5 would be nice if I could span volumes sanely across my (often changing) disks ...
If I needed RAID6, I would have to do with the smallest disk-size, but otherwise I don't see a problem.
So, what'd you make out of: [0:0:0:0] disk ATA SAMSUNG HD501LJ CR10 /dev/sda # .5T [1:0:0:0] disk ATA SAMSUNG HD204UI 1AQ1 /dev/sdb # 2T [2:0:0:0] disk ATA ST31500341AS CC1H /dev/sdc # 1.5T [3:0:0:0] disk ATA SAMSUNG HD203WI 1AN1 /dev/sdd # 2T [4:0:0:0] disk ATA SAMSUNG HD203WI 1AN1 /dev/sde # 2T [5:0:0:0] disk ATA SAMSUNG HD203WI 1AN1 /dev/sdf # 2T [8:0:0:0] disk ATA SAMSUNG HD204UI 1AQ1 /dev/sdg # 2T [9:0:0:0] disk ATA SAMSUNG HD204UI 1AQ1 /dev/sdh # 2Y [16:0:1:0] disk ATA WDC WD5000AVJB-6 05.0 /dev/sdi # .5T and that's rather recently, until a few months ago that was a merry mix of .5T, .75T, 1T, 1.5T and 2T ... Oh, and only /dev/sda (current system disk for a SUSE and a Gentoo), /dev/sdb (newest disk, prepped for 2 system-roots w/o home[1] to replace sda sometime), and /dev/sdg (one "home2" partition for basically replacable "home" stuff, one "data" partition). The rest is data partitions (one each disk). Oh, and expect disks being replaced by e.g. 3T drives soonish. -dnh PS: ask if you need help configuring freedos for running the Samsung FW-Update-Tool F4EG.exe for the HD204UI ... (I needed it only for 2 of those 3 disks, sda already came with the updated firmware). [1] lately I tend to use only / and /home partitions. -- Kiss me twice. I'm schizophrenic. -- from the BSD fortune file -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
David Haller wrote:
Hello,
sorry it took me so long to reply.
On Tue, 12 Apr 2011, Per Jessen wrote:
I would probably make a number of RAID1 arrays of equal-size disks, then use LVM for managing the space.
RAID1 as underlying RAID-Level would be to wasteful for my data.
It's not only about the data, it's also about availability and uptime. Maybe that's irrelevant to you, I don't know. Never mind, if CHF100 for a 2nd 2Tb drive is too much, and you're happy with a restore of yesterdays backup to the new drive (once you've bought it), RAID1 is wasteful, I agree.
RAID5 would be nice if I could span volumes sanely across my (often changing) disks ...
"span volumes"? Changing disks is only a practical matter, plus the time it takes to rebuild the array after a change.
If I needed RAID6, I would have to do with the smallest disk-size, but otherwise I don't see a problem.
So, what'd you make out of:
[0:0:0:0] disk ATA SAMSUNG HD501LJ CR10 /dev/sda # .5T [1:0:0:0] disk ATA SAMSUNG HD204UI 1AQ1 /dev/sdb # 2T [2:0:0:0] disk ATA ST31500341AS CC1H /dev/sdc # 1.5T [3:0:0:0] disk ATA SAMSUNG HD203WI 1AN1 /dev/sdd # 2T [4:0:0:0] disk ATA SAMSUNG HD203WI 1AN1 /dev/sde # 2T [5:0:0:0] disk ATA SAMSUNG HD203WI 1AN1 /dev/sdf # 2T [8:0:0:0] disk ATA SAMSUNG HD204UI 1AQ1 /dev/sdg # 2T [9:0:0:0] disk ATA SAMSUNG HD204UI 1AQ1 /dev/sdh # 2Y [16:0:1:0] disk ATA WDC WD5000AVJB-6 05.0 /dev/sdi # .5T
I would create a RAID6 on sdb, sdd, sde, sdf, sdg, sdh. That would give you an 8Tb array with dual drive redundancy. With 2Tb drives, recovering after a drive failure can be quite slow, giving plenty of time for a 2nd drive to fail. Your Samsung drives are relatively cheap desktop drives, so likely to fail exactly when the warranty expires :-( Alternatively, if dual-drive redundancy is too wasteful, a RAID5 over the same drives would give you 10Tb. Then I would probably run a RAID1 on sda and sdi for system etc etc.
and that's rather recently, until a few months ago that was a merry mix of .5T, .75T, 1T, 1.5T and 2T ... Oh, and only /dev/sda (current system disk for a SUSE and a Gentoo), /dev/sdb (newest disk, prepped for 2 system-roots w/o home[1] to replace sda sometime), and /dev/sdg (one "home2" partition for basically replacable "home" stuff, one "data" partition). The rest is data partitions (one each disk). Oh, and expect disks being replaced by e.g. 3T drives soonish.
With a forever changing setup like that, LVM sounds like just what you need. -- Per Jessen, Zürich (9.6°C) -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
David Haller wrote:
PS: ask if you need help configuring freedos for running the Samsung FW-Update-Tool F4EG.exe for the HD204UI ... (I needed it only for 2 of those 3 disks, sda already came with the updated firmware).
I'm curious. I thought one problem with the Samsung drives and fix was that the firmware version number didn't change when updated. So how do you tell that it already had the fix applied? Cheers, Dave -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Hello, On Tue, 19 Apr 2011, Dave Howorth wrote:
David Haller wrote:
PS: ask if you need help configuring freedos for running the Samsung FW-Update-Tool F4EG.exe for the HD204UI ... (I needed it only for 2 of those 3 disks, sda already came with the updated firmware).
I'm curious. I thought one problem with the Samsung drives and fix was that the firmware version number didn't change when updated. So how do you tell that it already had the fix applied?
heise wrote in the c't magazine (where I first read about the problem too), that drives with production dates after 2010/12 came with the updated firmware (and my third one has 2011/02 as production date). Oh, darn. Samsungs selling the HDD stuff to Seagate :((( -dnh -- It's so nice to be insane, no one asks you to explain! -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
David Haller wrote:
heise wrote in the c't magazine (where I first read about the problem too), that drives with production dates after 2010/12 came with the updated firmware (and my third one has 2011/02 as production date).
Thanks :)
Oh, darn. Samsungs selling the HDD stuff to Seagate :(((
Yes, need to look for another vendor yet again :( Cheers, Dave -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
David Haller said the following on 04/12/2011 11:35 AM:
I HAVE differently sized disks. How to have a sane RAID thingy across those without going bonkers? I'd be real happy if you'd come up with a good idea!
That's a very good point, David, and is the great shortcoming of RAID s RAID, as opposed to the "Raid" subsystems that are available from some commercial (read closed source) vendors. Being in a similar situation, I use LVM. I'm experimenting with Btrfs. RAID users keep one drive spare. In effect I do too, but its spread across all the other drives so I can make the best use of striping. I can also "mirror" file systems :-) My real problem is that many of my drives are older and I only have four slots on the motherboard and one us used by the CD/DVD drive. So new drives are going to be SATA, or I need to get a disk driver board. I've had couple of drives go bad. One was on a server that had been running for over 2 years when I shut it down to go on an extended vacation. On my return it crashed, irrecoverably ... as in head-gouging. I'm told that's a known failure mode if you don't shut down or "park the head" on some drives. The other was a slow decay. Increasing loss of sectors. Predicted death imminent. I added a drive and did a pvmove. The nice thing about a pvmove is that you can restart it if it is interrupted by a crash. Before you raise the matter, yes I do have a log-watcher and one of its patterns to look for is things that may be or become disk problems Things like smartd[3332]: Device: /dev/sda [SAT], 39 Currently unreadable (pending) sectors smartd[3332]: Device: /dev/sda [SAT], SMART Usage Attribute: 194 Temperature_Celsius changed from 157 to 171 -- "I think there is a world market for about five computers." Thomas J. Watson, chairman of the board of IBM, 1943 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Hello, sorry for my delay On Tue, 12 Apr 2011, Anton Aylward wrote:
David Haller said the following on 04/12/2011 11:35 AM:
I HAVE differently sized disks. How to have a sane RAID thingy across those without going bonkers? I'd be real happy if you'd come up with a good idea!
That's a very good point, David, and is the great shortcoming of RAID s RAID, as opposed to the "Raid" subsystems that are available from some commercial (read closed source) vendors.
I'm not so sure about the latter ;)
Being in a similar situation, I use LVM. I'm experimenting with Btrfs.
I like what zfs claims to do.
RAID users keep one drive spare. In effect I do too, but its spread across all the other drives so I can make the best use of striping. I can also "mirror" file systems :-)
I like it simple, best, if I could use a rescue-system with a hex-editor to repair a broken system[1]. LVM over RAID is just a "no-no" in my mind. What I need a backup of, I mirror via rsync locally to another drive and/or to a drive in the other box. The rest is redundant (at different levels) anyway, e.g. DVD .iso, dvdrip-rips, and resulting reencoded files.
My real problem is that many of my drives are older and I only have four slots on the motherboard and one us used by the CD/DVD drive. So new drives are going to be SATA, or I need to get a disk driver board.
Get a GA-770TA-UD3 if you're using AMD AM2+/AM3 CPUs ;) And yes, I know your dilemma, I've used a Athlon 500 on a MSI MS-6167 MoBo until last October or so (BIOS dated: 1999). I had 4 disks in that thing (since ~'00, last was 1120GB, IDE only), also a Promise Ultra 33 at times, and the PCI-SATA-Controller worked but hung up the box _hard_ when other PCI-traffic was occurring. I guess it were the ISA Sound- or SCSI-Card that confused the Controller (in the other box (the one with the MA770), that same SATA- Controller works nicely and reliably).
I've had couple of drives go bad.
Depending on the number of drives involved ... to be expected :(
One was on a server that had been running for over 2 years when I shut it down to go on an extended vacation. On my return it crashed, irrecoverably ... as in head-gouging. I'm told that's a known failure mode if you don't shut down or "park the head" on some drives.
*OUCH* I guess that's why SUSE runs a 'hdparm -y' on drives on shutdown since a while ago.
The other was a slow decay. Increasing loss of sectors.
I "phase out" disks once they report any defective sectors. They'll get used for "scrap space" and whatever, but not anything even remotely worth a backup.
Predicted death imminent. I added a drive and did a pvmove. The nice thing about a pvmove is that you can restart it if it is interrupted by a crash.
man rsync ;) Maybe not quite as efficient as pvmove.
Before you raise the matter, yes I do have a log-watcher and one of its patterns to look for is things that may be or become disk problems
Things like
smartd[3332]: Device: /dev/sda [SAT], 39 Currently unreadable (pending) sectors
*ouch* replace ASAP or use just as a temp-dev (somewhat like a RAM-Disk, but a bit more persistent, or something like that ;)
smartd[3332]: Device: /dev/sda [SAT], SMART Usage Attribute: 194 Temperature_Celsius changed from 157 to 171
I have a root-xterm running, with an 'tail -f' on /v/l/m (and more). -dnh [1] once you recreated an EPBR from scratch[2] from surrounding data (i.e. the MBR and the first logical partition and the following EPBR) with assembler and INT13 with DOS debug.com, you tend to like that "simplicity" ;) [2] was completely overwritten with 0xF6 by some partitioning (IIRC Cnegvgvba Zntvp) tool being confused -- So Linus, what are we doing tonight? The same thing we do every night Tux. Try to take over the world! -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
David Haller said the following on 04/18/2011 02:32 PM:
On Tue, 12 Apr 2011, Anton Aylward wrote:
David Haller said the following on 04/12/2011 11:35 AM:
I HAVE differently sized disks. How to have a sane RAID thingy across those without going bonkers? I'd be real happy if you'd come up with a good idea!
That's a very good point, David, and is the great shortcoming of RAID s RAID, as opposed to the "Raid" subsystems that are available from some commercial (read closed source) vendors.
I'm not so sure about the latter ;)
Neither am I :-) I've seen their explanations but still have this feeling that the salesdroids have a rubber chicken they are waving while your attention is misdirected.
Being in a similar situation, I use LVM. I'm experimenting with Btrfs.
I like what zfs claims to do.
So did I. But look at the licensing and look where Mason is working. http://thread.gmane.org/gmane.comp.file-systems.btrfs/2880 See also Alex Elsayed's comment in that thread Btrfs got to where it is so quickly b/c it is building on mature kernel features. For instance it’s using the same device mapping code for RAID and block abstraction that Linux’s software RAID and LVM are based on. Which is why it was merged into the mainline kernel so quickly.
RAID users keep one drive spare. In effect I do too, but its spread across all the other drives so I can make the best use of striping. I can also "mirror" file systems :-)
LVM over RAID is just a "no-no" in my mind.
For what I'm doing its not so much a "no-no" as a "Why?" I can do many RAID things with LVM - mirror, stripe- and more.
What I need a backup of, I mirror via rsync locally to another drive and/or to a drive in the other box.
I do that sometimes, too, but snapshot+DVD is my baseline. The snapshot is almost instantaneous so I can use it on live file systems in a way that rsync/tar/cpio can't deal with..
I've had couple of drives go bad.
Depending on the number of drives involved ... to be expected :(
One was on a server that had been running for over 2 years when I shut it down to go on an extended vacation. On my return it crashed, irrecoverably ... as in head-gouging. I'm told that's a known failure mode if you don't shut down or "park the head" on some drives.
*OUCH* I guess that's why SUSE runs a 'hdparm -y' on drives on shutdown since a while ago.
The other was a slow decay. Increasing loss of sectors.
I "phase out" disks once they report any defective sectors. They'll get used for "scrap space" and whatever, but not anything even remotely worth a backup.
Since modern drives remap defective sectors to the reserve area, what is your threshold? If I 'phased out' a drive when the first error was reported few drives would last 3 months! Bathtub means you get a few when its young -- that's life.
I have a root-xterm running, with an 'tail -f' on /v/l/m (and more).
So do I, on my log-server .... "h look, never mind you missed it..." So I have "swatch". -- Though force can protect in emergency, only justice, fairness, consideration and co-operation can finally lead men to the dawn of eternal peace. -- Dwight D. Eisenhower -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
$ grep 'sd.[0-9]$' /proc/partitions 8 17 31457280 sdb1 8 18 1048576 sdb2 8 19 31457280 sdb3 8 20 1 sdb4 8 21 1889550304 sdb5 8 1 5253223 sda1 8 2 1060290 sda2 8 3 31463302 sda3 8 4 1 sda4 8 5 5253223 sda5 8 6 31463271 sda6 8 7 413890596 sda7 8 49 1953514552 sdd1 8 65 1953513560 sde1 8 81 1953514552 sdf1 8 97 1044188 sdg1 8 98 10490448 sdg2 8 99 1941977362 sdg3 8 113 1953514552 sdh1 8 33 1465136001 sdc1 8 129 488384001 sdi1
How do you sanely handle this? Why so many partitions? It seems to me, this is precisely the situation where you'd want LVM. With it, you could create whatever partitions you want in the LVM and adjust them as needed. Also, if you have mulitple drives, you can have a RAID array and place the LVM on top of it. That way, you don't have to worry about drives failing, as you simply replace
David Haller wrote: them and rebuilt the RAID array. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Hello, sorry for my delay in answering. On Tue, 12 Apr 2011, James Knott wrote:
$ grep 'sd.[0-9]$' /proc/partitions 8 17 31457280 sdb1 8 18 1048576 sdb2 8 19 31457280 sdb3 8 20 1 sdb4 8 21 1889550304 sdb5 8 1 5253223 sda1 8 2 1060290 sda2 8 3 31463302 sda3 8 4 1 sda4 8 5 5253223 sda5 8 6 31463271 sda6 8 7 413890596 sda7 8 49 1953514552 sdd1 8 65 1953513560 sde1 8 81 1953514552 sdf1 8 97 1044188 sdg1 8 98 10490448 sdg2 8 99 1941977362 sdg3 8 113 1953514552 sdh1 8 33 1465136001 sdc1 8 129 488384001 sdi1
How do you sanely handle this? Why so many partitions? It seems to me, this is precisely the situation where you'd want LVM. With it, you could create whatever
David Haller wrote: partitions you want in the LVM and adjust them as needed. Also, if you have mulitple drives, you can have a RAID array and place the LVM on top of it. That way, you don't have to worry about drives failing, as you simply replace them and rebuilt the RAID array.
You need to hone your reading skills, i.e. focus on the letters between the "sd" and the numbers. That's: - 1 drive with 2 systems (each with a / and /home, data stuff is (sym-)linked into the homes), plus one "data" partition (sda) and a swap. - 1 drive prepped for 2 systems (/ only + a "data") and a swap (sdb) (and which is intended to migrate the sda systems to) - 1 drive with an extra "home2" (for misc stuff) and a swap (sdg) - all other drives: 1 partition/drive, formatted with 'mk2fs -j -T largefile -m 0 ...' How's that "so many" partitions? How else could you have ~13TiBi formatted capacity? -dnh -- I have always wished that my computer would be as easy to use as my telephone. My wish has come true. I no longer know how to use my telephone. -- Bjarne Stroustrup -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
David Haller said the following on 04/12/2011 09:04 AM:
Let's say I'd use LVM, what if e.g. /dev/sde shows defects (seen in the SMART data). Can LVM show me what dirs and files(!) I have on the PV(s) on that probably soon-failing drive?
I think you are asking the wrong question. If that drive is failing and and all files in any and all file systems in any and all LVs on any and all PVs on that drive are at risk. You don't need to know the individual file names. Thinking you have to backup/restore or move those files or file systems is "fixed partition" thinking. LVM thinking is tell the LVM system to stop using /dev/sde and to migrate the LVs (and hence the file systems and hence the files) to another PV. This is one reason why experienced users of LVM don't allocate all of the space to start with. We keep some around for snapshots and other contingencies, like this. Yes, you have 'scrap space'. But the thing about LVM is that it does this at a lower level, so you don't have to worry about re-mounting and playing involved games with symlinks, partition tables and such LVM is more KISS than all those fixed size partitions. If you really, really, really need it, LVM has the tools to tell you the mapping of a LV to disk sectors. Other file system tools can tell you what blocks in the file system a file uses. I saw on AIX tools that would tell you the answer to your question as it was posed, but then AIX *only* has LVM. As far as Murphy goes * All Flesh is Grass (Isaiah 40:6) * All Hardware is Rust and Sand * All Software is BitRot But for resilience and flexibility, using LVM beats out using fixed location and fixed size partitions. -- If you lie to the compiler, it will get its revenge. - Henry Spencer -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Hello, sorry for my delay ... On Tue, 12 Apr 2011, Anton Aylward wrote:
David Haller said the following on 04/12/2011 09:04 AM:
Let's say I'd use LVM, what if e.g. /dev/sde shows defects (seen in the SMART data). Can LVM show me what dirs and files(!) I have on the PV(s) on that probably soon-failing drive?
I think you are asking the wrong question.
Am I? What'd you do if you see a SMART failure for Sector 1234567 on device /dev/sdc? Wouldn't you want to replay the backup of the affected file _first_ (to another drive), so you have the file in duplicate again? Your backup _can_ fail too, and after Murphy, it will. c.f.: RAID: One more disk fails than can be recovered by the redundancy. -- Andreas Dau
If that drive is failing and and all files in any and all file systems in any and all LVs on any and all PVs on that drive are at risk. You don't need to know the individual file names.
Exactly my point! Which PVs and which LVs? Without RAID and LVM (and with ext2/3/4), you have _only one_ file (the one using that failed sector) affected until the drive actually dies. And (with debugfs) I can find out what file is affected. And of course, you replace the drive as soon as you can. But you still might want to reduplicate the (now) single ("current") instance of that file in the backup ASAP, so that there, again, is a redundancy between backup and live system. A file only in the backup is not a backup.
Thinking you have to backup/restore or move those files or file systems is "fixed partition" thinking. LVM thinking is tell the LVM system to stop using /dev/sde and to migrate the LVs (and hence the file systems and hence the files) to another PV.
And how do you know that the affected file will not be "migrated" to that "other" PV incorrectly? Will you get (and see) logs that that file caused an disk-IO-Error and that it is corrupted?
This is one reason why experienced users of LVM don't allocate all of the space to start with. We keep some around for snapshots and other contingencies, like this.
See above.
Yes, you have 'scrap space'. But the thing about LVM is that it does this at a lower level, so you don't have to worry about re-mounting and playing involved games with symlinks, partition tables and such
I don't need to. I choose to. And from an informed point of view, IMO.
LVM is more KISS than all those fixed size partitions.
Do you _REALLY_ understand LVM (and a possible underlying RAID) to the point to fix problems from a rescue system with - a hex-editor (e.g. vche or debug.com) - e2fsck [-b SUPERBLOCK] - debugfs (the latter two: or similar for other filesystems, if existant, else also with a hex-editor) I DO on my system!
If you really, really, really need it, LVM has the tools to tell you the mapping of a LV to disk sectors. Other file system tools can tell you what blocks in the file system a file uses. I saw on AIX tools that would tell you the answer to your question as it was posed, but then AIX *only* has LVM.
I still avoid it.
As far as Murphy goes * All Flesh is Grass (Isaiah 40:6) * All Hardware is Rust and Sand * All Software is BitRot
See above.
But for resilience and flexibility, using LVM beats out using fixed location and fixed size partitions.
I'm quite happy with my setup. I know and understand symlinks and 'mount --bind' intrinsically, with DOS-partitions down to the bit. And: symlinks work over NFS (once the NFS-stuff is mounted). ~10T of my stuff is in the other box, mounted via NFS. YMMV, of course! -dnh -- Tupperware wurde erfunden, damit man die Lebensmittel erst noch 14 Tage aufbewahren kann, bevor man sie wegwirft. -- Holger Hirschfeld -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
David Haller said the following on 04/18/2011 03:45 PM:
Hello,
sorry for my delay ...
On Tue, 12 Apr 2011, Anton Aylward wrote:
David Haller said the following on 04/12/2011 09:04 AM:
Let's say I'd use LVM, what if e.g. /dev/sde shows defects (seen in the SMART data). Can LVM show me what dirs and files(!) I have on the PV(s) on that probably soon-failing drive?
I think you are asking the wrong question.
Am I? What'd you do if you see a SMART failure for Sector 1234567 on device /dev/sdc?
Nothing. Its also telling me its remapped the drive. By the time you see that message the drive firmware has see soft errors and tried correcting them, seen it happen a few times and become unhappy with that sector and remapped it to one of the 'spare' ones for recovery. Back in the days of SCO XENIX on the PDP-11 I wrote an error correcting disk river and saw it made redundant by on-board drivers for SCSI and later; *sigh* it was a crazy hack, having to run on the kernel stack and deferring the the logical disk access while did CRC calculations between interrupts and if required did the remapping. http://www.travelnotes.de/california/silicon/oldpdp.htm My laptop's smartd does in fact report some bum sectors after all these years. Laptops are not as amenable to RAID as desktops and servers. Does it bother me? Yes and no. No, I expect it after a few years. Every disk has its bathtub. Yes, it tells me to get a new laptop :-)
Wouldn't you want to replay the backup of the affected file _first_ (to another drive), so you have the file in duplicate again? Your backup _can_ fail too, and after Murphy, it will. c.f.:
We can play *that* mindgame into infinite regression. We can also argue about whether you bother backing up "system software" that can be reinstalled. We can argue about how you run integrity checks.
RAID: One more disk fails than can be recovered by the redundancy. -- Andreas Dau
If that drive is failing and and all files in any and all file systems in any and all LVs on any and all PVs on that drive are at risk. You don't need to know the individual file names.
Exactly my point! Which PVs and which LVs? Without RAID and LVM (and with ext2/3/4), you have _only one_ file (the one using that failed sector) affected until the drive actually dies. And (with debugfs) I can find out what file is affected.
I think you have an incomplete understanding of how LVM works. You might as well make the same argument about any system of partitioning. All LVM is doing is introducing one more level of indirection. That applies to access as well as debug. With 'hard sectors' you'd still have to build a complete map between files and physical sectors. With 'soft sectors' of LVM you build a complete map between files and logical sectors in just the same way -- then using the tables that the LVM tools supply, map those to the logical sectors. Personally I think the mapping of the files is the big job :-) Has it been done? Years ago I did something like this in an attempt to come up with a de-fragmentation application; I had it working for some special cases but became frustrated and showed the the backers that a backup-mkfs-restore was quicker and more general. I hear that some people are working on sector-level deduplication -- finding common sector images between files and 'linking them', and developing a file system to support that. From your POV having more than one file using a sector that goes bad is a high risk, but their scanning tools for conversion might be of interest. Yes its all computationally heavy, but you're the one that
And of course, you replace the drive as soon as you can.
In an ideal world. In a real world you don't do a Chicken Little on the first error report. Check your drive specs. and don't forget that unless you have a mainframe-grade environmental enclosure for your drives, the chassis, cabling and power supply are all as unreliable as the drives. My laptop's drive has lasted longer than two batteries, a screen and a touchpad. On my file/mail server, the one that sees most use, I've had two motherboards die, and the video of a third, closely followed by the internal Ethernet. Disk drives have not been a problem :-)
But you still might want to reduplicate the (now) single ("current") instance of that file in the backup ASAP, so that there, again, is a redundancy between backup and live system. A file only in the backup is not a backup.
That is one reason I use LVM. It makes backups - snapshots - so much easier.
Thinking you have to backup/restore or move those files or file systems is "fixed partition" thinking. LVM thinking is tell the LVM system to stop using /dev/sde and to migrate the LVs (and hence the file systems and hence the files) to another PV.
And how do you know that the affected file will not be "migrated" to that "other" PV incorrectly? Will you get (and see) logs that that file caused an disk-IO-Error and that it is corrupted?
That gets back to indefinite regression argument ... is it real or is it Emulex? Which is the correct one of two mirrored drives? We could go on forever with this game.
LVM is more KISS than all those fixed size partitions.
Do you _REALLY_ understand LVM (and a possible underlying RAID) to the point to fix problems from a rescue system with
- a hex-editor (e.g. vche or debug.com) - e2fsck [-b SUPERBLOCK] - debugfs
No, because that's not the way to fix things. Its only *A* way. You're asking a question that presupposes your answer and your POV is the correct one, and that things are structured in a way that is suitable to those tools. I recall one seeing one kindergarten child ask another "What would you rather do, eat a box of chalk or drink a jar of paint?" RAID you say? Can I pick PAID1 - Mirrors? Can I then put the two drives in another box, mount the file systems and then do a side-by-side tree-walk to see where files differ? No need for he editors ...
I DO on my system!
Good for you. I don't.
But for resilience and flexibility, using LVM beats out using fixed location and fixed size partitions.
I'm quite happy with my setup. I know and understand symlinks and 'mount --bind' intrinsically, with DOS-partitions down to the bit.
And there's nothing to stop you using that on top of LVM
And: symlinks work over NFS (once the NFS-stuff is mounted). ~10T of my stuff is in the other box, mounted via NFS.
YMMV, of course!
Mine is driven by two things. The first is that LVM doesn't care wow big the drives are. Well neither does Btrfs, if you want to have just "/" and 'pretend to have mounted partitions, that may be a good choice as well. I've been playing with Btrfs on one machine and it seems OK. I'll watch how it evolves; it may do very well for SSD, particularly on a laptop in place of LVM. Because LVM doesn't care about drive size I can mirror and stripe in ways I can't with traditional RAID. The second is backup. LVM lets me take snapshots. It make disk-to-disk-to-tape or in my case disk-to-disk-to-DVD very easy. The easier a backup is the less likely one is to skip it. And those DVDs have usable file systems so making finding an old rile and restoring it easy. I can do other things with LVM. I can plug in an external drive, make it part of the group, snapshot to it. Fast backup! Faster than a file system walk. And the backup is a usable file system. Obsessing about disk failure when using a system that has so many other failure modes that statistics show are equally likely or more likely strikes me as foolish. In the 'real world' I've been using AIX since it was first released; I was using LVM and RAID on large AIX multiprocessor rigs long before LVM was available or Linux and was impressed with it. -- The emphasis should be on "why" we do a job - W. Edwards Deming -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
David Haller wrote:
If that drive is failing and and all files in any and all file systems in any and all LVs on any and all PVs on that drive are at risk. You don't need to know the individual file names.
Exactly my point! Which PVs and which LVs? Without RAID and LVM (and with ext2/3/4), you have _only one_ file (the one using that failed sector) affected until the drive actually dies. And (with debugfs) I can find out what file is affected.
Why go through all that when you can just instruct LVM to stop using the failing drive? -- Per Jessen, Zürich (9.4°C) -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Per Jessen said the following on 04/19/2011 02:12 AM:
David Haller wrote:
If that drive is failing and and all files in any and all file systems in any and all LVs on any and all PVs on that drive are at risk. You don't need to know the individual file names.
Exactly my point! Which PVs and which LVs? Without RAID and LVM (and with ext2/3/4), you have _only one_ file (the one using that failed sector) affected until the drive actually dies. And (with debugfs) I can find out what file is affected.
Why go through all that when you can just instruct LVM to stop using the failing drive?
I suggested that, David wasn't happy with it. David seems to want file level granuality. If the drive is dying, I don't see the point. -- "It is impossible for a man to begin to learn what he thinks he knows". -- Epictetus -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On 4/9/2011 8:08 AM, Wolfgang Mueller wrote:
Yesterday I bought a new laptop and installed 11.4.
As it is long ago when I installed OpenSuse last time, I followed the suggestion to create only three Linux partitions:
Win7 167 GB (shrinked by the Installation CD) Swap 2 GB / 20 GB /home 260 GB (the remainder)
Normally I would have created a much larger number of Linux partitions, for instance:
Win7 167 GB (shrinked by the Installation CD, as above) Swap 2 GB / 10 GB /usr 20 GB /opt 10 GB /var 5 GB /tmp 5 GB /boot 0.09 GB /srv 60 GB /data 60 GB /local 60 GB /home 52 GB
Does that make sense? Or is it only a waste of space, since the partitions cannot be filled completely?
Thanks in advance for your suggestions.
Best regards, Wolfgang
Is it better to have a pile of 55-gallon drums or a swimming pool? Is it better to have a single tool handle with 50 bits that fit in the end, or 50 dedicated tools? Is it better to wear jeans, or a wool suit, or tights? The answer is whatever you need, you should do. There are many reasons to make a given directory into it's own filesystem, and many reasons not to. I could say "just make one fs, no /boot, no swap" and be perfectly right, for one scenario. And it would be the worst thing in the world for another scenario. I can't say do one big fs, because I don't know what you're actually doing with your box, so, with one big fs, /data or /home or /srv or /tmp may suddenly grow one day and bring the box down. I also can't say make a bunch of separate fs's. Maybe /data is some mission-critical business app that must never die, and the only point of the whole box is just to run that app, and even if /data one day grows beyond expectations and threatens the entire box, those extra weeks, or even extra hours of uptime might be more important than saving the rest of the box from going down by making that app die when it's artificially limited fs fills up. -- bkw -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
participants (19)
-
Anton Aylward -
auxsvr＠gmail.com -
Brian K. White -
C -
Ciro Iriarte -
Dave Howorth -
David C. Rankin -
David Haller -
Don Raboud -
Felix Miata -
Hans Witvliet -
James Knott -
jsa -
Ken Schneider - openSUSE -
Myrosia Dzikovska -
Per Jessen -
Rodney Baker -
wolf python london -
Wolfgang Mueller