#1010514 lvm2: Snapshot generates Buffer I/O error ... async read message during boot

Package:
lvm2
Source:
lvm2
Description:
Linux Logical Volume Manager
Submitter:
Vincas Dargis
Date:
2022-05-03 09:18:04 UTC
Severity:
normal
#1010514#5
Date:
2022-05-03 09:11:47 UTC
From:
To:
Dear Maintainer,

We have two similar machines working as database master & replica. They both
use same HDD drives and Areca ARC-1261 HW raid controllers, with same LVM
volume groups set up.

I've noticed that during boot, and before rebooting, both of these machines
produce error for the SAME block, but only when LVM snapshots exists:

```
# salt -L master.domain.com,replica.domain.com cmd.run "zgrep 'Buffer I/O error' /var/log/kern*"
replica.domain.com:
    /var/log/kern.log:May  3 10:44:59 replica kernel: [3114072.061835] Buffer I/O error on dev dm-5, logical block 861061104, async page read
    /var/log/kern.log:May  3 10:54:16 replica kernel: [    2.544139] Buffer I/O error on dev dm-5, logical block 861061104, async page read
    /var/log/kern.log:May  3 10:54:16 replica kernel: [    6.653298] Buffer I/O error on dev dm-5, logical block 861061104, async page read
    /var/log/kern.log.1:Apr  6 11:33:24 replica kernel: [784199.317321] Buffer I/O error on dev dm-5, logical block 861061104, async page read
    /var/log/kern.log.1:Apr  6 11:34:31 replica kernel: [784266.011713] Buffer I/O error on dev dm-5, logical block 861061104, async page read
    /var/log/kern.log.3.gz:Mar 28 09:39:26 replica kernel: [1551875.380127] Buffer I/O error on dev dm-5, logical block 861061104, async page read
    /var/log/kern.log.3.gz:Mar 28 09:43:27 replica kernel: [    2.565393] Buffer I/O error on dev dm-5, logical block 861061104, async page read
    /var/log/kern.log.3.gz:Mar 28 09:43:27 replica kernel: [    6.635266] Buffer I/O error on dev dm-5, logical block 861061104, async page read
master.domain.com:
    /var/log/kern.log:May  3 10:43:59 master kernel: [3070183.772497] Buffer I/O error on dev dm-6, logical block 861061104, async page read
    /var/log/kern.log.4.gz:Mar 28 10:16:18 master kernel: [905062.938666] Buffer I/O error on dev dm-6, logical block 861061104, async page read
    /var/log/kern.log.4.gz:Mar 28 21:56:28 master kernel: [    1.570582] Buffer I/O error on dev dm-6, logical block 861061104, async page read
    /var/log/kern.log.4.gz:Mar 28 21:56:28 master kernel: [    5.298972] Buffer I/O error on dev dm-6, logical block 861061104, async page read
```

After removing snapshot and rebooting machine the error no longer appear.

Oldest error was found on February (as far I could see in backups), probably
during some big database upgrade when snapshots are usually used:

```
kern.log.1:Feb 17 08:04:18 replica kernel: [    1.470778] Buffer I/O error on dev dm-3, logical block 861061104, async page read
```

So I am not sure if this started after some recent kernel upgrade, for example.

We did not discover any "real" issues (data corruption, instability, etc), only
this log message. Not sure how critical it is.