Bug#1032067: linux-image-5.10.0-19-amd64: recurring btrfs errors and warning related to ibm websphere mq operating files
Control: tags -1 + moreinfo
Hi,
On Mon, Feb 27, 2023 at 11:49:49AM +0100, IT wrote:
> Package: src:linux
> Version: 5.10.149-2
> Severity: normal
> X-Debbugs-Cc: it@bsi.sm
>
> Dear Maintainer,
>
> on a server with IBM WebSphere MQ, BTRFS returns these messages:
>
> BTRFS warning (device vda2): csum failed root 256 ino 647789 off 9109504 csum 0x2d57cc4b expected csum 0x58676c1e mirror 1
> BTRFS error (device vda2): bdev /dev/vda2 errs: wr 0, rd 0, flush 0, corrupt 2869, gen 0
> BTRFS warning (device vda2): csum failed root 256 ino 647789 off 9109504 csum 0x2d57cc4b expected csum 0x58676c1e mirror 1
> BTRFS error (device vda2): bdev /dev/vda2 errs: wr 0, rd 0, flush 0, corrupt 2870, gen 0
> BTRFS warning (device vda2): csum failed root 256 ino 647789 off 9109504 csum 0x2d57cc4b expected csum 0x58676c1e mirror 1
> BTRFS error (device vda2): bdev /dev/vda2 errs: wr 0, rd 0, flush 0, corrupt 2871, gen 0
> BTRFS warning (device vda2): csum failed root 256 ino 647789 off 9109504 csum 0x2d57cc4b expected csum 0x58676c1e mirror 1
> BTRFS error (device vda2): bdev /dev/vda2 errs: wr 0, rd 0, flush 0, corrupt 2872, gen 0
> BTRFS warning (device vda2): csum failed root 256 ino 647789 off 9109504 csum 0x2d57cc4b expected csum 0x58676c1e mirror 1
> BTRFS error (device vda2): bdev /dev/vda2 errs: wr 0, rd 0, flush 0, corrupt 2873, gen 0
>
> These are the last ten lines from dmesg, the rest is filled with similar
> lines. The complete dmesg refers to these inodes/files (some characters
> are redacted with ?):
>
> $ ls -il $(..)
> 647486 -rw-rw---- 1 mqm mqm 16785408 23 feb 11.31 IBM/MQ/data/log/??????0/active/S0000001.LOG
> 647487 -rw-rw---- 1 mqm mqm 16785408 27 feb 10.19 IBM/MQ/data/log/??????0/active/S0000002.LOG
> 647789 -rw-rw---- 1 mqm mqm 16785408 27 feb 07.30 IBM/MQ/data/log/??????1/active/S0000000.LOG
> 649095 -rw-rw---- 1 mqm mqm 16785408 25 gen 09.10 IBM/MQ/data/log/??????0/active/S0000000.LOG
> 650011 -rw-rw---- 1 mqm mqm 16785408 27 feb 10.23 IBM/MQ/data/log/??????P/active/S0000000.LOG
> 650012 -rw-rw---- 1 mqm mqm 16785408 8 feb 16.49 IBM/MQ/data/log/??????P/active/S0000001.LOG
> 650013 -rw-rw---- 1 mqm mqm 16785408 20 feb 07.45 IBM/MQ/data/log/??????P/active/S0000002.LOG
>
> An example directory containing these files:
>
> $ ls -il IBM/MQ/data/log/??????0/active/
> 647485 -rw-rw---- 1 mqm mqm 16785408 20 feb 07.36 S0000000.LOG
> 647486 -rw-rw---- 1 mqm mqm 16785408 23 feb 11.31 S0000001.LOG
> 647487 -rw-rw---- 1 mqm mqm 16785408 27 feb 11.17 S0000002.LOG
>
> This is a relatively new install of Debian 11 running in a virtual
> machine. BTRFS is mounted with:
>
> $ grep btrfs /etc/fstab
> UUID=? / btrfs subvol=@rootfs,lazytime 0 0
>
> Previously the filesystem was mounted with compression enabled (with
> mount option compress=zstd:1) and WebSphere MQ crashed with messages
> like this (some values are redacted with ?):
>
> ----- amqrcsia.c : 820 --------------------------------------------------------
> 01/08/2023 10:21:45 PM - Process(7948.81351) User(mqm) Program(amqrmppa)
> Host(?) Installation(?)
> VRMF(9.3.0.0) QMgr(?)
> Time(2023-01-08T21:21:45.811Z)
> CommentInsert1(?)
> CommentInsert2(7948)
> CommentInsert3(? (?))
>
> AMQ9999E: Channel '?' to host '? (?)' ended
> abnormally.
>
> EXPLANATION:
> The channel program running under process ID 7948 for channel
> '?' ended abnormally. The host name is '? (?)'; in
> some cases the host name cannot be determined and so is shown as '????'.
> ACTION:
> Look at previous error messages for the channel program in the error logs to
> determine the cause of the failure. Note that this message can be excluded
> completely or suppressed by tuning the "ExcludeMessage" or "SuppressMessage"
> attributes under the "QMErrorLog" stanza in qm.ini. Further information can be
> found in the System Administration Guide.
> ----- amqrmrsa.c : 632 --------------------------------------------------------
>
> With compression disabled the kernel writes the messages reported at the
> beginning but WebSphere MQ works. While worrying the program works as
> expected. I might be wrong but the affected files seems to be some sort
> of log similar in purpose to PostgreSQL WAL files (both binary).
>
> Any hints or ideas are appreciated.
As 5.10.149-2 is not the most recent kernel in bullseye, please try in
any case as well the newest one, 5.10.162-1, this might not resolve
your problem but gives you at least the correct baseline.
Regards,
Salvatore
Reply to: