http://bugzilla.novell.com/show_bug.cgi?id=1121221
http://bugzilla.novell.com/show_bug.cgi?id=1121221#c3
--- Comment #3 from Neil Brown ---
I've been having a bit of a look at this.
The messages "getattr ERROR: Request couldn't be completed in time" means the
an NFS getattr request got an NFS4ERR_DELAY response, which means that the
server needed to get some sort of authentication information from user-space
and it took too long (which is a second or 2 I think). This isn't fatal, but
might hint at a problem.
One way that NFSv4.x+KRB5 is different from the other combinations is that it
requests the kernel nfs server to ask rpc.idmapd to map between user-names and
uids.
If rpc.idmapd were blocked by something, that would explain the slow mount, and
the NFs4ERR_DELAY errors.
It would also explain occasional delays in "find" if (and only if) different
files in the tree were sometimes owned by different users.
You could try killing rpc.idmapd (systemctl stop nfs-idmapd) and then run
rpc.idmapd -f -vvv
(on the NFS server)
and watch the output while trying to mount the filesystem, or do the 'find'.
--
You are receiving this mail because:
You are on the CC list for the bug.