I just tried converting a FS from reiser to ext4 and got a ridiculous 'out of space' message - it turned out that the 'lots-a-small-files' meant I ran out of inodes long, long before I ran out of data space. Does reiser allocate inodes dynamically? Anyway, I wondered if there was a tool that that walked a file system and told you the distribution of file sizes, so as to help you figure out the inode vs space ratio. -- An expert is someone who knows some of the worst mistakes that can be made in his subject and how to avoid them. --Werner Heisenberg -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Hello, On Fri, 08 Jun 2012, Anton Aylward wrote:
I just tried converting a FS from reiser to ext4 and got a ridiculous 'out of space' message - it turned out that the 'lots-a-small-files' meant I ran out of inodes long, long before I ran out of data space.
Does reiser allocate inodes dynamically?
Yes. That's why you get no useful info using 'df -i'.
Anyway, I wondered if there was a tool that that walked a file system and told you the distribution of file sizes, so as to help you figure out the inode vs space ratio.
reiserfsck outputs the number of dirs and files, so with the size of the FS you can calculate the ratio. If in doubt you could use 'mke2fs -T news' with an inode for every 4096 bytes (which is the usual blocksize). HTH, -dnh -- SYNOPSIS glimpse - [almost all letters] pattern -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 2012-06-08 22:07, Anton Aylward wrote:
I just tried converting a FS from reiser to ext4 and got a ridiculous 'out of space' message - it turned out that the 'lots-a-small-files' meant I ran out of inodes long, long before I ran out of data space.
Does reiser allocate inodes dynamically?
Yes. You could also use xfs. - -- Cheers / Saludos, Carlos E. R. (from 11.4 x86_64 "Celadon" at Telcontar) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.16 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAk/ScA0ACgkQIvFNjefEBxrDBwCgpLZpCYVatIXXNDgOG0GKZewr w9wAoK4hvLK4ITqIYzk08+Tg+PYJkfAm =jLMp -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 06/08/2012 04:35 PM, Carlos E. R. wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 2012-06-08 22:07, Anton Aylward wrote:
I just tried converting a FS from reiser to ext4 and got a ridiculous 'out of space' message - it turned out that the 'lots-a-small-files' meant I ran out of inodes long, long before I ran out of data space.
Does reiser allocate inodes dynamically? Yes.
You could also use xfs.
That's is exactly why I gave up on ext3 and ext4 because I've got too many small files. And for whatever reason, I found reiserfs 100 x more reliable than ext3. I've been using reiserfs for over 3 years and have NOT had to do a complete rebuild of the fs yet (except when I do upgrades - I always rebuild for upgrades - for me, it just works better) Duaine -- Duaine Hechler Piano, Player Piano, Pump Organ Tuning, Servicing& Rebuilding Reed Organ Society Member Florissant, MO 63034 (314) 838-5587 dahechler@att.net www.hechlerpianoandorgan.com -- Home& Business user of Linux - 11 years -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Fri, 2012-06-08 at 16:45 -0500, Duaine Hechler wrote:
On 06/08/2012 04:35 PM, Carlos E. R. wrote:
On 2012-06-08 22:07, Anton Aylward wrote:
I just tried converting a FS from reiser to ext4 and got a ridiculous 'out of space' message - it turned out that the 'lots-a-small-files' meant I ran out of inodes long, long before I ran out of data space. Does reiser allocate inodes dynamically? Yes. You could also use xfs. That's is exactly why I gave up on ext3 and ext4 because I've got too many small files. And for whatever reason, I found reiserfs 100 x more reliable than ext3. I've been using reiserfs for over 3 years and have NOT had to do a complete rebuild of the fs yet (except when I do upgrades - I always rebuild for upgrades - for me, it just works better)
I've been using ext3/4 for that long and haven't had to do a filesystem. These days, in this use case, I'd go with btrfs. It's ready. But XFS is perfectly adequate. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 06/08/2012 05:19 PM, Adam Tauno Williams wrote:
On Fri, 2012-06-08 at 16:45 -0500, Duaine Hechler wrote:
On 06/08/2012 04:35 PM, Carlos E. R. wrote:
On 2012-06-08 22:07, Anton Aylward wrote:
I just tried converting a FS from reiser to ext4 and got a ridiculous 'out of space' message - it turned out that the 'lots-a-small-files' meant I ran out of inodes long, long before I ran out of data space. Does reiser allocate inodes dynamically? Yes. You could also use xfs. That's is exactly why I gave up on ext3 and ext4 because I've got too many small files. And for whatever reason, I found reiserfs 100 x more reliable than ext3. I've been using reiserfs for over 3 years and have NOT had to do a complete rebuild of the fs yet (except when I do upgrades - I always rebuild for upgrades - for me, it just works better) I've been using ext3/4 for that long and haven't had to do a filesystem.
These days, in this use case, I'd go with btrfs. It's ready.
But XFS is perfectly adequate.
So, does btrfs replay some sort of journal if the system locks up or crashes. Lately, the flash plugin has been crashing left and right and even sometimes locks up my whole computer, where I have to hit the reset button to get back up. Thanks, Duaine -- Duaine Hechler Piano, Player Piano, Pump Organ Tuning, Servicing& Rebuilding Reed Organ Society Member Florissant, MO 63034 (314) 838-5587 dahechler@att.net www.hechlerpianoandorgan.com -- Home& Business user of Linux - 11 years -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Sat, 09 Jun 2012 10:27:39 Duaine Hechler wrote:
... Lately, the flash plugin has been crashing left and right and even sometimes locks up my whole computer, where I have to hit the reset button to get back up.
Last year I had a similar issue, easily repeatable by using speedtest.net's heavily flashy test site. In my case I found that overly enthusiatic default motherboard settings were the cause - the TPU/EPU overclocking switches were shipped in the on position. Forum posting here: http://forums.opensuse.org/english/get-technical-help-here/applications/4689... -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Adam Tauno Williams said the following on 06/08/2012 06:19 PM:
These days, in this use case, I'd go with btrfs. It's ready.
I tried BtrFS and wasn't happy with it. I should like it; its more to do with ReiserFS than ext4, but ... Somehow the idea of a big FS that is managed makes me nervous. Is there a proper FSCK yet? I though a FCKS was O(2) or worse, and one big FS with internal subvolume logical mounts seems to defeat the purpose of having smaller FS and running the FSCK on each in parallel. The Phoronix tests seem to indicate Btrfs vs ext4 is very OS level dependant. http://www.phoronix.com/scan.php?page=news_item&px=MTExMjg -- "Opportunity is missed by most people because it is dressed in overalls and looks like work." -- Thomas A. Edison -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Anton Aylward said the following on 06/08/2012 04:07 PM:
I just tried converting a FS from reiser to ext4
... because when I shut down, sometimes I get a kernel crash when un-mounting a resiserFS. Oh it cleans up OK on the ext boot. Perhaps I have an old version? -- If a little knowledge is dangerous, where is the man who has so much as to be out of danger. Thomas H. Huxley -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 2012-06-09 00:54, Anton Aylward wrote:
Anton Aylward said the following on 06/08/2012 04:07 PM:
I just tried converting a FS from reiser to ext4
... because when I shut down, sometimes I get a kernel crash when un-mounting a resiserFS.
You should bugzilla it. - -- Cheers / Saludos, Carlos E. R. (from 11.4 x86_64 "Celadon" at Telcontar) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.16 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAk/Sj2MACgkQIvFNjefEBxrLFwCgspObblliXmYumpmrDPvU5en3 ckUAn09ok9tU3ggL0lqXUG/rPkCwgUf7 =EJA1 -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Carlos E. R. said the following on 06/08/2012 07:48 PM:
On 2012-06-09 00:54, Anton Aylward wrote:
Anton Aylward said the following on 06/08/2012 04:07 PM:
I just tried converting a FS from reiser to ext4
... because when I shut down, sometimes I get a kernel crash when un-mounting a resiserFS.
You should bugzilla it.
Yea right. it isn't consistent; doesn't happen when I unmounted the dozen or so FS by hand and I can't take a snapshot cost it not logged - its s shutdown crash at the console. I was wondering of anyone has any thoughts abnout a historic reason. The FS were set up long ago under mandriva. Converted to suse ... I also wonder ... there are FS mounted on /home which is a mounted FS. Is the unmount -a in the shutdown smart enough to unwind properly or does it read though /etc/fstab from the begining I have /--+ +- /home --+ +-- /anton --+ + -- /Documents + -- /Mail --+ + -- /current + -- /archive Is unmount smart enough to work 'bottom up'? -- Nothing excites a magical particle like meeting itself coming the other way. -- _The Science of Discworld_ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 2012-06-09 02:53, Anton Aylward wrote:
I also wonder ... there are FS mounted on /home which is a mounted FS. Is the unmount -a in the shutdown smart enough to unwind properly or does it read though /etc/fstab from the begining
I have /--+ +- /home --+ +-- /anton --+ + -- /Documents + -- /Mail --+ + -- /current + -- /archive
Is unmount smart enough to work 'bottom up'?
I think so, but also fstab has to be in the right order during mount, so I suppose on umount it uses the inverse order. - -- Cheers / Saludos, Carlos E. R. (from 11.4 x86_64 "Celadon" at Telcontar) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.16 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAk/VL00ACgkQIvFNjefEBxq9+QCgo+SeYicDCtYqUGXK05EVE5oE BAMAoK8lxHoPFqxZ5FM36GiOHi+8hrIg =ZK18 -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Carlos E. R. said the following on 06/10/2012 07:35 PM:
On 2012-06-09 02:53, Anton Aylward wrote:
I also wonder ... there are FS mounted on /home which is a mounted FS. Is the unmount -a in the shutdown smart enough to unwind properly or does it read though /etc/fstab from the begining
I have /--+ +- /home --+ +-- /anton --+ + -- /Documents + -- /Mail --+ + -- /current + -- /archive
Is unmount smart enough to work 'bottom up'?
I think so, but also fstab has to be in the right order during mount, so I suppose on umount it uses the inverse order.
While its a reasonable assumption, I'm not sure. The only thing I can find for sure is that the /etc/fstab has a field that controls the order in which the fsck is run. I realise that with systemd I will be able to hand-craft scripts that create dependencies, so that ~anton/PDF will be dependent on ~anton being mounted, which will in term be dependent on /home being mounted. But this is just 11.4, no systemd. -- We'll give you your money back if not satisfied, but we do suspect that we will be quite satisfied with your money. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Anton Aylward said the following on 06/10/2012 08:41 PM:
Carlos E. R. said the following on 06/10/2012 07:35 PM:
On 2012-06-09 02:53, Anton Aylward wrote:
I also wonder ... there are FS mounted on /home which is a mounted FS. Is the unmount -a in the shutdown smart enough to unwind properly or does it read though /etc/fstab from the begining
I have /--+ +- /home --+ +-- /anton --+ + -- /Documents + -- /Mail --+ + -- /current + -- /archive
Is unmount smart enough to work 'bottom up'?
I think so, but also fstab has to be in the right order during mount, so I suppose on umount it uses the inverse order.
While its a reasonable assumption, I'm not sure. The only thing I can find for sure is that the /etc/fstab has a field that controls the order in which the fsck is run.
To be fair, the ftab man page says The order of records in fstab is important because fsck(8), mount(8), and umount(8) sequentially iterate through fstab doing their thing. But that implies umount works though the same way that mount does, beginning to end, which is crazy. I suppose you could buqqer things up by setting the 6th field to be '2' and all others to be '1' so that the root FS gets fsck'd last :-) -- I suspect that, over time, all bureaucratic processes decay into cargo cults unless regularly challenged by a hostile reality. - Alan Rocker, 23-Nov-2011 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 2012-06-11 02:41, Anton Aylward wrote:
Carlos E. R. said the following on 06/10/2012 07:35 PM:
On 2012-06-09 02:53, Anton Aylward wrote:
Is unmount smart enough to work 'bottom up'?
I think so, but also fstab has to be in the right order during mount, so I suppose on umount it uses the inverse order.
While its a reasonable assumption, I'm not sure. The only thing I can find for sure is that the /etc/fstab has a field that controls the order in which the fsck is run.
I know for sure that mounting happens in the order that the lines are read in fstab, because if you revert the lines, they are not mounted - at least with systemv.
I realise that with systemd I will be able to hand-craft scripts that create dependencies, so that ~anton/PDF will be dependent on ~anton being mounted, which will in term be dependent on /home being mounted.
But this is just 11.4, no systemd.
Fortunately... - -- Cheers / Saludos, Carlos E. R. (from 11.4 x86_64 "Celadon" at Telcontar) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.16 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAk/VQaQACgkQIvFNjefEBxqqsgCgodxW+TwUrSDg7LrTcpw3OnLV /q4AoKJ0xlDbDKyLtnEV3+1xWfvgLoIj =E5Cb -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Carlos E. R. said the following on 06/10/2012 08:53 PM:
But this is just 11.4, no systemd.
Fortunately...
Oh, I don't know. I have another machine here that has been running Fedora with systemd and it runs quite painlessly and reliably, none of the problems I've seen reported with systemd on openSuse. If it wasn't for the fact I hate the admin model of Fedora, yum and its stupid behaviour, all even worse than mandriva!, I'd give up and convert. Even moving though F15, F16, F17 has been painless. -- Ignorance feeds on ignorance. Science phobia is contagious. --Carl Sagan, The Demon-Haunted World -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 2012-06-11 03:35, Anton Aylward wrote:
Carlos E. R. said the following on 06/10/2012 08:53 PM:
But this is just 11.4, no systemd.
Fortunately...
Oh, I don't know. I have another machine here that has been running Fedora with systemd and it runs quite painlessly and reliably, none of the problems I've seen reported with systemd on openSuse.
The order of things is unpredictable, some things do not work right - I have seen recent posts about the problems with nfs mounted on boot via fstab, does not work. You have to instead mount via autofs.
If it wasn't for the fact I hate the admin model of Fedora, yum and its stupid behaviour, all even worse than mandriva!, I'd give up and convert. Even moving though F15, F16, F17 has been painless.
To each his own :-) - -- Cheers / Saludos, Carlos E. R. (from 11.4 x86_64 "Celadon" at Telcontar) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.16 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAk/VTVUACgkQIvFNjefEBxo7dQCgqmifwXoakYoNAWsb3pE6gerP SHYAoJTjcxJeeod9hkfFucB9PvrbGUbF =wbCa -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Fri, 08 Jun 2012 20:53:56 -0400
Anton Aylward
I have /--+ +- /home --+ +-- /anton --+ + -- /Documents + -- /Mail --+ + -- /current + -- /archive
Is unmount smart enough to work 'bottom up'?
It should be, otherwise there will be no clean shutdown for /home . -- Regards, Rajko -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Anton Aylward wrote:
I just tried converting a FS from reiser to ext4 and got a ridiculous 'out of space' message - it turned out that the 'lots-a-small-files' meant I ran out of inodes long, long before I ran out of data space.
FWIW, I use Reiser for the reason of lots of small files. I have tried both ext4 and xfs instead but both gave worse performance. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Dave Howorth said the following on 06/11/2012 06:16 AM:
Anton Aylward wrote:
I just tried converting a FS from reiser to ext4 and got a ridiculous 'out of space' message - it turned out that the 'lots-a-small-files' meant I ran out of inodes long, long before I ran out of data space.
FWIW, I use Reiser for the reason of lots of small files. I have tried both ext4 and xfs instead but both gave worse performance.
This is what I am finding ... -- Passive acceptance of the teacher's wisdom is easy to most boys and girls. It involves no effort of independent thought, and seems rational because the teacher knows more than his pupils; it is moreover the way to win the favour of the teacher unless he is a very exceptional man. Yet the habit of passive acceptance is a disastrous one in later life. It causes man to seek and to accept a leader, and to accept as a leader whoever is established in that position. Bertrand Russell (1872 - 1970) -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Anton Aylward wrote:
Dave Howorth said the following on 06/11/2012 06:16 AM:
Anton Aylward wrote:
I just tried converting a FS from reiser to ext4 and got a ridiculous 'out of space' message - it turned out that the 'lots-a-small-files' meant I ran out of inodes long, long before I ran out of data space. FWIW, I use Reiser for the reason of lots of small files. I have tried both ext4 and xfs instead but both gave worse performance.
This is what I am finding ...
FYI, there might be another factor, if you use RAID. I just saw this comment by Dave Chinner on the XFS list: "The reason, I'd suggest, is that you've chosen the wrong RAID volume type for your workload. Small random file read and write workloads like news and mail spoolers are IOPS intensive workloads and do not play well with RAID5/6. RAID5/6 really only work well for large files with sequential access patterns - you need to use RAID1/10 for IOPS intensive workloads because they don't suffer from the RMW cycle problem that RAID5/6 has for small writes. The iostat output will help clarify whether this is really the problem or not..." Luckily for me, I do use RAID10 on the XFS filesystem I do have, not because I'm clever but because the hardware I thought would do RAID6 turned out not to! So I've been content with mdadm RAID10 instead. Lucky escape :) -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
participants (8)
-
Adam Tauno Williams -
Anton Aylward -
Carlos E. R. -
Dave Howorth -
David Haller -
Duaine Hechler -
michael＠actrix.gen.nz -
Rajko