Gentoo Websites Logo
Go to: Gentoo Home Documentation Forums Lists Bugs Planet Store Wiki Get Gentoo!
Bug 9459 - Possible Gentoo-sources-r9/r10 Reiserfs corruption
Summary: Possible Gentoo-sources-r9/r10 Reiserfs corruption
Status: VERIFIED FIXED
Alias: None
Product: Gentoo Linux
Classification: Unclassified
Component: [OLD] Core system (show other bugs)
Hardware: x86 Linux
: High major (vote)
Assignee: Brandon Low (RETIRED)
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2002-10-21 17:26 UTC by Samuel Greenfeld
Modified: 2003-08-07 09:13 UTC (History)
3 users (show)

See Also:
Package list:
Runtime testing required: ---


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Samuel Greenfeld 2002-10-21 17:26:49 UTC
I am having corruption issues with gentoo-sources-r9 and -r10.  While the exact
cause has yet to be pinpointed, the suspicious is that if reiserfs is shut down
incorrectly, the journal replay seems to possibly truncate and/or delete files
that were written *well* before the abnormal reboot.  

So far, I have lost oddball man pages (that end up being truncated) and a major
X library (libXaw.so IIRC). Recompiling the owning package of the destroyed item
brings it back without any visable incident.

The issue might be due to tails being cut off, or something like that. The
resiserfs partition fstab entry is 
 dev/hda4               /               reiserfs        defaults 0 1

I am currently running acpi-sources 2.4.20-r5 (acpi disabled), since that
includes the 2.4.20-pre8 patch, which removes some not-well-tested resierfs code.
Comment 1 Brandon Low (RETIRED) gentoo-dev 2002-10-21 18:45:17 UTC
hmm... I may be having similar problems on MY system, I'll look deeper into it, 
ASAP, and if necessary push out a new gentoo-sources sooner than planned.
Comment 2 Samuel Greenfeld 2002-10-21 20:37:47 UTC
   Also of note for the record:   1. The corruption cases I have seen involve only a small number of files being  corrupted, making this issue hard to track down.  Truncation is only suspected because viewing the corrupted files shows them to be quite short.   2. It is _not_ known if acpi-sources-r5 (*) fixes this problem or not.  So far, I have not had a hard crash while running this kernel to "experimentally" see if this is the case.     It is, however, known from the generic 2.4.20-pre changelog that there are reiserfs changes/potential fixes in the -pre8 patch acpi-sources applies. (Whether or not the maintainers want to move up to what is now -pre11 is their own decision :)  (*) With ACPI disabled; not by choice but by my BIOS' ACPI table being blacklisted.  Strangely, the 2.4.x generic ACPI doesn't have any issues, but that's a whole other story... 
Comment 3 Don Seiler (RETIRED) gentoo-dev 2002-12-23 13:25:32 UTC
So ... should I go with vanilla sources?
Comment 4 Brandon Low (RETIRED) gentoo-dev 2002-12-31 18:23:52 UTC
I still haven't been able to duplicate this on -r9... my issues were on -r10 only.  

Please test the latest lolo-sources available in gentoo-stable, they rock the
kazbah (are good).  lolo-sources are entering the Release Candidate stage, so
expect a release of gentoo-sources sometime soon :)
Comment 5 Brandon Low (RETIRED) gentoo-dev 2003-01-09 15:09:01 UTC
let me know.
Comment 6 John Davis (zhen) (RETIRED) gentoo-dev 2003-04-04 01:19:18 UTC
db fix
Comment 7 John Davis (zhen) (RETIRED) gentoo-dev 2003-04-04 01:24:18 UTC
db fix
Comment 8 Martin Holzer (RETIRED) gentoo-dev 2003-08-07 09:13:02 UTC
closed