[Gluster-users] No healing, errno 22

Zenon Panoussis oracle at provocation.net
Wed Mar 17 16:53:05 UTC 2021

>> replicate-0: performing entry selfheal on 94aefa13-9828-49e5-9bac-6f70453c100f

> Does this gfid correspond to the same directory path as last time?

No, it's one of the "two unrelated directories" that
I mentioned in some previous post. Both directories
exist on the volume mount and contain files that can
be listed. The third gfid is probably a file.

> Wonder if you can attach gdb the glustershd process at 
> the function client4_0_mkdir() 

Ehum, after 20 years as a linux professional, I still
haven't come around to learning to use gdb. I thought
OK, I'll give it a shot now, but CentOS hasn't published
any gluster debuginfo since gluster v6.0:


I tried strace on glfs_shdheal however, and I get hundreds
and hundreds of

22000 futex(0x55c87bb00f7c, FUTEX_WAIT_BITSET_PRIVATE, 4473, {8652954, 543829556}, ffffffff) = -1 ETIMEDOUT (Connection timed out)
22000 <... futex resumed> )             = -1 ETIMEDOUT (Connection timed out)
22000 <... futex resumed> )             = -1 ETIMEDOUT (Connection timed out)

All three nodes were connected and could ping each-other,
so I don't know what connection kept timing out.

More information about the Gluster-users mailing list