Replies: 9 comments 23 replies
-
Is mergerfs running when that happens? |
Beta Was this translation helpful? Give feedback.
-
Remount failed. $ systemctl restart srv-7614e796\\x2db9da\\x2d42db\\x2db1cf\\x2d0b73d7d494b1.mount
Job failed. See "journalctl -xe" for details.
$ journalctl -u srv-7614e796\\x2db9da\\x2d42db\\x2db1cf\\x2d0b73d7d494b1.mount
-- Journal begins at Wed 2023-02-08 12:34:27 CST, ends at Wed 2023-02-08 22:44:09 CST. --
Feb 08 22:42:52 linas systemd[1]: Unmounting MergerFS mount for main...
Feb 08 22:42:52 linas umount[78216]: umount: /srv/7614e796-b9da-42db-b1cf-0b73d7d494b1: target is busy.
Feb 08 22:42:52 linas systemd[1]: srv-7614e796\x2db9da\x2d42db\x2db1cf\x2d0b73d7d494b1.mount: Mount process exited, code=exited, status=32/n/a
Feb 08 22:42:52 linas systemd[1]: Failed unmounting MergerFS mount for main.
Feb 08 22:44:09 linas systemd[1]: Unmounting MergerFS mount for main...
Feb 08 22:44:09 linas umount[78321]: umount: /srv/7614e796-b9da-42db-b1cf-0b73d7d494b1: target is busy.
Feb 08 22:44:09 linas systemd[1]: srv-7614e796\x2db9da\x2d42db\x2db1cf\x2d0b73d7d494b1.mount: Mount process exited, code=exited, status=32/n/a
Feb 08 22:44:09 linas systemd[1]: Failed unmounting MergerFS mount for main. |
Beta Was this translation helpful? Give feedback.
-
As it mentions the mount is busy so it can't be umounted. If the process is running then it hasn't crashed and you can do some straces as described in the docs to provide me with details. mergerfs is mostly just a proxy. It isn't returning many errors itself. I need a strace from mergerfs to see if/what is actually responding with the error. |
Beta Was this translation helpful? Give feedback.
-
Sorry, I don't konw how to use strace in right way. strace -fvTtt -s 256 -p 848 -o /tmp/mergerfs.strace.txt I collected 500 lines. Hope you can get something valuable.
|
Beta Was this translation helpful? Give feedback.
-
strace -fvTtt -s 256 -p 844 -o /tmp/mergerfs.strace.txt
strace -fvTtt -s 256 -o /tmp/app.strace.txt ls /srv/7614e796-b9da-42db-b1cf-0b73d7d494b1 /tmp/app.strace.txt
/tmp/mergerfs.strace.txt
|
Beta Was this translation helpful? Give feedback.
-
2.33 still happens. |
Beta Was this translation helpful? Give feedback.
-
I have write a script to monitor mergerfs with strace: #!/bin/bash
exitcode=0
while [[ $exitcode -eq 0 ]]
do
echo "start strace at $(date +%s)"
timeout 5 strace -fvTtt -s 256 -p `ps -ef | grep " mergerfs " | grep -v grep | awk '{print $2}'` -o /tmp/mergerfs.strace.txt
ls /srv/7614e796-b9da-42db-b1cf-0b73d7d494b1
exitcode=$?
done It would do:
So if it was broken within 5 seconds, I can get the strace log. Here is the strace log: https://gist.github.com/libook/46c14721d04b313f6311717f4f4f931b |
Beta Was this translation helpful? Give feedback.
-
Seems that I'm getting the same error. Almost everyday my mergerfs mount is getting I/O error.
If i'm stopping the NFS-Server, unmounting/remounting /mnt2/storage clears the issue. There is a constant traffic on the mount ±100MB/s 24/7. There are no issues with the drives. Thanks |
Beta Was this translation helpful? Give feedback.
-
Anyone having the issue: try the new release with the |
Beta Was this translation helpful? Give feedback.
-
Describe
Using the mount point every morning will prompt an
Input/output error
.To Reproduce
I tried to
cd
to the mount point.$ cd /srv/7614e796-b9da-42db-b1cf-0b73d7d494b1 cd: Input/output error: /srv/7614e796-b9da-42db-b1cf-0b73d7d494b1
It showed me the error message.
Check the systemd configuration.
Check 3 origins of the mergerfs.
They are all working.
Remount it.
It back to working.
System information:
df -h
lsblk -f
Additional context
Sorry, I don't know much about how to collect logs from the file system. If you need any additional information, you can tell me how to do it.
Beta Was this translation helpful? Give feedback.
All reactions