Hm, I see alot of these in the first brick log:
2021-03-25 16:26:46.194101] E [MSGID: 113118]
[posix-helpers.c:1535:posix_gfid_heal] 0-MRIData-posix: Fresh file:
/gluster1/data/.shard/fae3b4e3-f2d4-46bb-871a-84f1f0a9030d.1905 [No such file
or directory]
[2021-03-25 16:27:25.614158] E [MSGID:
Sorry for the delay.
Here is a text file with the basic information of the gluster:
https://filebin.net/j9yvkyo3vjkcozy9/glusterfs-setup.txt?t=t4tmsjyg
Here are the 3 brick log files:
1. https://filebin.net/j9yvkyo3vjkcozy9/gluster1-data.log-20210328.gz?t=t4tmsjyg
2.
This one looks interesting:
/usr/lib64/glusterfs/8.4/xlator/features/shard.so(+0x65bd) [0x7f176fdd45bd] )
0-MRIData-shard: Failed to get trusted.glusterfs.shard.file-size for
67a6-b81f-48da-9359-2e2260b253c5
Maybe it's time to add gluster-devel to the loop.
Best Regards,Strahil Nikolov
Sure.
* old log, before downgrade: https://pastebin.com/ZUd3dG36 (when you
see references to 'fff' errors is me trying to create a file named
'fff' from command line)
* new log: https://pastebin.com/MXm1z14Z
Thanks,
On 3/29/21 1:07 PM, Strahil Nikolov wrote:
Consider sharing the logs
OK, I downgraded glusterfs on my Ovirt server to 7.9 (got the RPMs from
the CentOS mirror and did a "dnf --allowerasing downgrade ") and I
no longer have gluster errors.
Thanks for the help.
--
As a result of Coronavirus-related precautions, NYU and the Center for
Brain Imaging operations