| View previous topic :: View next topic |
| Author |
Message |
MrUlterior
Guru
Joined: 22 Mar 2005
Posts: 511
Location: Switzerland
|
 Posted: Mon Sep 04, 2006 8:48 am Post subject: reiserfs: more than 48hours for --rebuild-tree on 1TB FS? |
 |
|
Well, as the topic suggests, I have a 1.08TB RAID1 array (mixed IDE & SATA); reccently I noticed some errors after a crash. So I disabled all services, removed the array from its mountpoint and ran:
| Code: | | fsck.reiserfs --fix-fixable /dev/vg/data --yes |
The output of this indicated that there was some corruption and I should run:
| Code: | | fsck.reiserfs --rebuildtree /dev/vg/data --yes |
Which I did. I had a sense of foreboding when I saw the "this will take a long time" warning. However, now it's been going for more than 8 hours and has only done about 20%.
It still has 156821024 blocks to go, at approx 1363 blocks/second, this will take another 32 hours ....
Is this normal? Am I going to have a usable FS after this?
_________________
Misanthropy 2.0 - enough hate to go around
|
|
| Back to top |
|
 |
frostschutz
Advocate
Joined: 22 Feb 2005
Posts: 2977
Location: Germany
|
| Posted: Mon Sep 04, 2006 9:50 am Post subject: |
|
|
Chances are that it'll get faster as more blocks get done (less blocks to consider). Check back in 1-2 hours and see if it is doing more blocks/second then. Other than that, you'll just have to wait and see ... of course this is taking way too long and you don't know wether the result will be readable at all, but hey ... it's reiserfs  |
|
| Back to top |
|
|
MrUlterior
Guru
Joined: 22 Mar 2005
Posts: 511
Location: Switzerland
|
| Posted: Mon Sep 04, 2006 3:49 pm Post subject: |
|
|
Well, 8 hours later and it's now just over 60% -- which still leaves at least another 12 hours to go. It appears you are correct 
Let's just hope I have a usable filesystem after this!
_________________
Misanthropy 2.0 - enough hate to go around
|
|
| Back to top |
|
|
GNUtoo
Veteran
Joined: 05 May 2005
Posts: 1919
|
| Posted: Mon Sep 04, 2006 9:56 pm Post subject: |
|
|
yes that is a problem
that´s one of the reason was a meeting about filesystem in linux kernel
with person like linus tovald mabe andrew morton etc... |
|
| Back to top |
|
|
MrUlterior
Guru
Joined: 22 Mar 2005
Posts: 511
Location: Switzerland
|
| Posted: Mon Sep 04, 2006 10:17 pm Post subject: |
|
|
| Code: | | reiserfsck finished at Tue Sep 5 00:10:50 2006 |
| Quote: | Sep 5 00:11:33 xenophobe ReiserFS: dm-1: found reiserfs format "3.6" with standard journal
Sep 5 00:12:09 xenophobe ReiserFS: dm-1: using ordered data mode
Sep 5 00:12:09 xenophobe ReiserFS: dm-1: journal params: device dm-1, size 8192, journal first block 18, max trans len 1024, max batch 900, max commit age 30, max trans age 30
Sep 5 00:12:09 xenophobe ReiserFS: dm-1: checking transaction log (dm-1)
Sep 5 00:12:09 xenophobe ReiserFS: dm-1: Using r5 hash to sort names
|
The filesystem mounted, there're a few files in "lost+found", but it looks good
I guess I'll be sticking with reiserfs on my LVM, if it recovers so well after such a hard crash, it seems reliable enough for my purposes -- even if it is slow!
_________________
Misanthropy 2.0 - enough hate to go around
|
|
| Back to top |
|
|
GNUtoo
Veteran
Joined: 05 May 2005
Posts: 1919
|
| Posted: Tue Sep 05, 2006 12:19 am Post subject: |
|
|
| personaly with some data loss in reiserfs i will switch to ext3 |
|
| Back to top |
|
|
MrUlterior
Guru
Joined: 22 Mar 2005
Posts: 511
Location: Switzerland
|
| Posted: Tue Sep 05, 2006 8:19 am Post subject: |
|
|
| new_to_non_X86 wrote: | | personaly with some data loss in reiserfs i will switch to ext3 |
I think you misunderstand how the data was lost. I had a crash, after which I ran reisfer.fsck --fix-fixable --yes on the array this replayed the journal, fixed some things and aborted as there were other errors, it also advised me to run --rebuild-tree.
Foolishly, I ignored this, mounted the filesystem and continued working with the FS. A few weeks later, I noticed I was unable to delete some files and the system tended to kernel panic when I decompressed large files anywhere on the FS's mountpoint.
It's only then I begun the process described in this post. And the only files lost seem to be the contents of the .tar.bz2 I was decompressing and a movie that was copying via NFS when the kernel panic occurred. If I really wanted to recover these, I'd just have to rename the files in "lost+found"; however I don't so I just nuked them.
In my experience with ext3, it would have fared the same or worse in these conditions. What makes you think it'd do better?
_________________
Misanthropy 2.0 - enough hate to go around
|
|
| Back to top |
|
|
GNUtoo
Veteran
Joined: 05 May 2005
Posts: 1919
|
| Posted: Tue Sep 05, 2006 11:30 am Post subject: |
|
|
| MrUlterior wrote: | | new_to_non_X86 wrote: | | personaly with some data loss in reiserfs i will switch to ext3 |
I think you misunderstand how the data was lost. I had a crash, after which I ran reisfer.fsck --fix-fixable --yes on the array this replayed the journal, fixed some things and aborted as there were other errors, it also advised me to run --rebuild-tree.
Foolishly, I ignored this, mounted the filesystem and continued working with the FS. A few weeks later, I noticed I was unable to delete some files and the system tended to kernel panic when I decompressed large files anywhere on the FS's mountpoint.
It's only then I begun the process described in this post. And the only files lost seem to be the contents of the .tar.bz2 I was decompressing and a movie that was copying via NFS when the kernel panic occurred. If I really wanted to recover these, I'd just have to rename the files in "lost+found"; however I don't so I just nuked them.
In my experience with ext3, it would have fared the same or worse in these conditions. What makes you think it'd do better? |
ok |
|
| Back to top |
|
|
|
Other Things Gentoo
