You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Type | Version/Name
Linux
Distribution Name Kubuntu
Distribution Version 14.04.6 with lots of updates
Linux Kernel 4.19.133-ck1
Architecture x86_64
ZFS Version 0.8.4
SPL Version idem
Describe the problem you're observing
This system is a "root-on-zfs" system that I have been running under VirtualBox off an actual harddisk, via VBox's "raw disk access". This never gave issues with kernels up to and including 4.14.23 . Importing the pool under 4.19.133 has led to the following report:
pool: bolaLNXBak
state: ONLINE
status: One or more devices has experienced an error resulting in data
corruption. Applications may be affected.
action: Restore the file in question if possible. Otherwise restore the
entire pool from backup.
see: http://zfsonlinux.org/msg/ZFS-8000-8A
scan: resilvered 222G in 0 days 04:48:06 with 0 errors on Tue Jul 14 04:04:08 2020
config:
NAME STATE READ WRITE CKSUM SLOW
bolaLNXBak ONLINE 0 0 0 -
ata-VBOX_HARDDISK_VBe9d3fab6-73a1dc23-part6 ONLINE 0 0 0 0 (trim unsupported)
errors: Permanent errors have been detected in the following files:
<metadata>:<0x0>
<metadata>:<0x1>
<metadata>:<0x1dd>
bolaLNXBak/opt/MacPorts:<0x0>
bolaLNXBak/opt/MacPorts:<0x42606>
bolaLNXBak/opt/MacPorts:<0x44108>
bolaLNXBak/opt/MacPorts:<0xf209>
bolaLNXBak/opt/MacPorts:<0x4260e>
bolaLNXBak/opt/MacPorts:<0x42614>
bolaLNXBak/opt/MacPorts:<0x44117>
bolaLNXBak/opt/MacPorts:<0x4652f>
bolaLNXBak/opt/MacPorts:<0x44130>
bolaLNXBak/opt/MacPorts:<0x23a41>
bolaLNXBak/opt/MacPorts:<0x44143>
bolaLNXBak/opt/MacPorts:<0x44164>
bolaLNXBak/opt/MacPorts:<0x44264>
bolaLNXBak/opt/MacPorts:<0x6b369>
bolaLNXBak/opt/MacPorts:<0x2628d>
bolaLNXBak/opt/MacPorts:<0x45ebc>
bolaLNXBak/opt/MacPorts:<0x442c0>
bolaLNXBak/opt/MacPorts:<0x442d3>
bolaLNXBak/opt/MacPorts:<0x45edd>
bolaLNXBak/opt/MacPorts:<0xf0>
bolaLNXBak/opt/MacPorts:<0x440ff>
bolaLNXBak/opt/MacPorts:<0x463ff>
Many of the hex numbered entries were first listed as actual files that supposedly had an error - but which had the exact same md5sum as a valid copy on a different system. Simply restoring the file from backup (or cloning it and them moving the clone over the original) replaced the named error with the useless (to me) listings above.
Describe how to reproduce the problem
The first errors appeared out of the blue and concerned only directories that AFAIK were never written to. I've also seen new errors appear when restoring packages that had supposedly faulty files.
The 3 entries above aside, the errors are all in a single dataset. I noticed that sync was disabled in that dataset; I have not yet seen new errors appear since re-enabling sync.
My kernel has the Con Kolivas patches (as all my kernel builds) but also a submitted patch to make zswab use B-Trees (probably irrelevant) and the NixOS patch posted by @mskarbek (#8836 (comment)_) to re-enable the export of 2 SIMD functions. Could this be the culprit?
I know ZFS tests for the existence of all functions it needs, but could it have an additional hardcoded assumption that 4.19 doesn't support using SIMD operations?
Include any warning/errors/backtraces from the system logs
None.
I'll be testing some more with a different (root on ~zfs) to see if I can observe other cases of supposed errors.
The text was updated successfully, but these errors were encountered:
This issue has been automatically marked as "stale" because it has not had any activity for a while. It will be closed in 90 days if no further activity occurs. Thank you for your contributions.
System information
Type | Version/Name
Linux
Distribution Name Kubuntu
Distribution Version 14.04.6 with lots of updates
Linux Kernel 4.19.133-ck1
Architecture x86_64
ZFS Version 0.8.4
SPL Version idem
Describe the problem you're observing
This system is a "root-on-zfs" system that I have been running under VirtualBox off an actual harddisk, via VBox's "raw disk access". This never gave issues with kernels up to and including 4.14.23 . Importing the pool under 4.19.133 has led to the following report:
Many of the hex numbered entries were first listed as actual files that supposedly had an error - but which had the exact same md5sum as a valid copy on a different system. Simply restoring the file from backup (or cloning it and them moving the clone over the original) replaced the named error with the useless (to me) listings above.
Describe how to reproduce the problem
The first errors appeared out of the blue and concerned only directories that AFAIK were never written to. I've also seen new errors appear when restoring packages that had supposedly faulty files.
The 3 entries above aside, the errors are all in a single dataset. I noticed that sync was disabled in that dataset; I have not yet seen new errors appear since re-enabling sync.
My kernel has the Con Kolivas patches (as all my kernel builds) but also a submitted patch to make zswab use B-Trees (probably irrelevant) and the NixOS patch posted by @mskarbek (#8836 (comment)_) to re-enable the export of 2 SIMD functions. Could this be the culprit?
I know ZFS tests for the existence of all functions it needs, but could it have an additional hardcoded assumption that 4.19 doesn't support using SIMD operations?
Include any warning/errors/backtraces from the system logs
None.
I'll be testing some more with a different (root on ~zfs) to see if I can observe other cases of supposed errors.
The text was updated successfully, but these errors were encountered: