On 06/28/2016 02:27 AM, Ian Kent wrote:
Can you describe again how this fits together?
Sure. Please let me know if I'm not clear. Here's my /etc/auto.master:
/nfs program:/etc/auto.nfs nobind
/home program:/etc/auto.home --mode=751
/etc/auto.nfs does some sanity checks, then does:
echo -fstype=nfs4,noatime,nosuid,_netdev,soft,intr,timeo=1000 $1:/
/etc/auto.home looks up in an internal database on what NFS server the
account is, then does:
echo -fstype=bind :/nfs/httpX/$1
So /nfs and /home are entirely handled by autofs. There's no NFS or user
mount handled by something else.
On the NFS servers, /home is a standard filesystem (XFS), here's the
/etc/exports:
/home 1.2.3.4 (rw,async,insecure,no_subtree_check,no_root_squash,fsid=0)
On my server that deadlocks the most, all users are on the same NFS
server (named http12).
At this point a full debug log would probably answer most of my questions.
I've just enabled debug logs, I'll send them to you as soon as I have a
new deadlock.
Thanks again.
--
Cyril B.
--
To unsubscribe from this list: send the line "unsubscribe autofs" in
