Bug 882550

Summary: autofs timeouts get really streeeeched
Product: [Fedora] Fedora Reporter: Michal Jaegermann <michal>
Component: autofsAssignee: Ian Kent <ikent>
Status: CLOSED CURRENTRELEASE QA Contact: Fedora Extras Quality Assurance <extras-qa>
Severity: unspecified Docs Contact:
Priority: unspecified    
Version: 17CC: ikent
Target Milestone: ---Keywords: Reopened
Target Release: ---   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2012-12-12 04:30:45 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Attachments:
Description Flags
Patch - add timeout option description to man page none

Description Michal Jaegermann 2012-12-01 20:10:36 UTC 
Description of problem:

This is the scenario: autofs is running on Fedora 17 client running 3.6.7-4.fc17.x86_64 kernel patched with a patch from attachment 654705 [details], bug 874372, in order for NFS mounts to work. /etc/auto.master has the following in it:

        /net	-hosts -timeout=60

The following timeline is from a remote ssh login and there are no other logins on a client so a desktop stuff is not running.  Here we go:

# date
Sat Dec  1 12:22:46 MST 2012
# ls /net/zeno/home/
  ( ... list requested directory ... )
# mount | egrep 'autofs|nfs'
systemd-1 on /proc/sys/fs/binfmt_misc type autofs (rw,relatime,fd=27,pgrp=1,timeout=300,minproto=5,maxproto=5,direct)
/etc/auto.misc on /misc type autofs (rw,relatime,fd=7,pgrp=1289,timeout=300,minproto=5,maxproto=5,indirect)
-hosts on /net type autofs (rw,relatime,fd=13,pgrp=1289,timeout=300,minproto=5,maxproto=5,indirect)
zeno:/ on /net/zeno type nfs (rw,nosuid,nodev,relatime,vers=3,rsize=524288,wsize=524288,namlen=255,hard,proto=tcp,timeo=600,retrans=2,sec=sys,mountaddr=192.168.x.y,mountvers=3,mountport=20048,mountproto=udp,local_lock=none,addr=192.168.x.y)
-hosts on /net/zeno/boot type autofs (rw,relatime,fd=13,pgrp=1289,timeout=300,minproto=5,maxproto=5,offset)
-hosts on /net/zeno/home type autofs (rw,relatime,fd=13,pgrp=1289,timeout=300,minproto=5,maxproto=5,offset)
-hosts on /net/zeno/opt type autofs (rw,relatime,fd=13,pgrp=1289,timeout=300,minproto=5,maxproto=5,offset)
-hosts on /net/zeno/usr type autofs (rw,relatime,fd=13,pgrp=1289,timeout=300,minproto=5,maxproto=5,offset)
-hosts on /net/zeno/var type autofs (rw,relatime,fd=13,pgrp=1289,timeout=300,minproto=5,maxproto=5,offset)
zeno:/home on /net/zeno/home type nfs (rw,nosuid,nodev,relatime,vers=3,rsize=524288,wsize=524288,namlen=255,hard,proto=tcp,timeo=600,retrans=2,sec=sys,mountaddr=192.168.x.y,mountvers=3,mountport=20048,mountproto=udp,local_lock=none,addr=192.168.x.y)
-hosts on /net/zeno/home/spare type autofs (rw,relatime,fd=13,pgrp=1289,timeout=300,minproto=5,maxproto=5,offset)
# date
Sat Dec  1 12:27:02 MST 2012
# mount | grep nfs
zeno:/ on /net/zeno type nfs (rw,nosuid,nodev,relatime,vers=3,rsize=524288,wsize=524288,namlen=255,hard,proto=tcp,timeo=600,retrans=2,sec=sys,mountaddr=192.168.x.y,mountvers=3,mountport=20048,mountproto=udp,local_lock=none,addr=192.168.x.y)
zeno:/home on /net/zeno/home type nfs (rw,nosuid,nodev,relatime,vers=3,rsize=524288,wsize=524288,namlen=255,hard,proto=tcp,timeo=600,retrans=2,sec=sys,mountaddr=192.168.x.y,mountvers=3,mountport=20048,mountproto=udp,local_lock=none,addr=192.168.x.y)
# date
Sat Dec  1 12:32:16 MST 2012
# mount | grep nfs
zeno:/ on /net/zeno type nfs (rw,nosuid,nodev,relatime,vers=3,rsize=524288,wsize=524288,namlen=255,hard,proto=tcp,timeo=600,retrans=2,sec=sys,mountaddr=192.168.x.y,mountvers=3,mountport=20048,mountproto=udp,local_lock=none,addr=192.168.x.y)
# date
Sat Dec  1 12:33:22 MST 2012
# mount | egrep 'autofs|nfs'
systemd-1 on /proc/sys/fs/binfmt_misc type autofs (rw,relatime,fd=27,pgrp=1,timeout=300,minproto=5,maxproto=5,direct)
/etc/auto.misc on /misc type autofs (rw,relatime,fd=7,pgrp=1289,timeout=300,minproto=5,maxproto=5,indirect)
-hosts on /net type autofs (rw,relatime,fd=13,pgrp=1289,timeout=300,minproto=5,maxproto=5,indirect)

In other words it took close to ten minutes before only zeno:/ remained mounted; after that it indeed another minute, roughly, was needed before all nfs mounts expired.

I did not peek constantly into mounts so times recorded are not precise. That still shows orders of magnitude.

Version-Release number of selected component (if applicable):
autofs-5.0.6-22.fc17.x86_64

How reproducible:
all the time

Additional info:

Here is the same scenario but this time from Fedora 16 client with
autofs-5.0.6-8.fc16.x86_64, with the same configuration,  and running 3.6.7-4.fc16.x86_64 kernel (also patched the same way as above).

# date
Sat Dec  1 11:37:55 MST 2012
# ls /net/zeno/home
  ( ... list requested directory ... )
# mount | egrep 'autofs|nfs'
systemd-1 on /proc/sys/fs/binfmt_misc type autofs (rw,relatime,fd=28,pgrp=1,timeout=300,minproto=5,maxproto=5,direct)
sunrpc on /var/lib/nfs/rpc_pipefs type rpc_pipefs (rw,relatime)
-hosts on /net type autofs (rw,relatime,fd=6,pgrp=1060,timeout=60,minproto=5,maxproto=5,indirect)
zeno:/ on /net/zeno type nfs (rw,nosuid,nodev,relatime,vers=3,rsize=524288,wsize=524288,namlen=255,hard,proto=tcp,timeo=600,retrans=2,sec=sys,mountaddr=192.168.x.y,mountvers=3,mountport=20048,mountproto=udp,local_lock=none,addr=192.168.x.y)
-hosts on /net/zeno/boot type autofs (rw,relatime,fd=6,pgrp=1060,timeout=60,minproto=5,maxproto=5,offset)
-hosts on /net/zeno/home type autofs (rw,relatime,fd=6,pgrp=1060,timeout=60,minproto=5,maxproto=5,offset)
-hosts on /net/zeno/opt type autofs (rw,relatime,fd=6,pgrp=1060,timeout=60,minproto=5,maxproto=5,offset)
-hosts on /net/zeno/usr type autofs (rw,relatime,fd=6,pgrp=1060,timeout=60,minproto=5,maxproto=5,offset)
-hosts on /net/zeno/var type autofs (rw,relatime,fd=6,pgrp=1060,timeout=60,minproto=5,maxproto=5,offset)
zeno:/home on /net/zeno/home type nfs (rw,nosuid,nodev,relatime,vers=3,rsize=524288,wsize=524288,namlen=255,hard,proto=tcp,timeo=600,retrans=2,sec=sys,mountaddr=192.168.x.y,mountvers=3,mountport=20048,mountproto=udp,local_lock=none,addr=192.168.x.y)
-hosts on /net/zeno/home/spare type autofs (rw,relatime,fd=6,pgrp=1060,timeout=60,minproto=5,maxproto=5,offset)
# date
Sat Dec  1 11:39:23 MST 2012
# mount | grep nfs
sunrpc on /var/lib/nfs/rpc_pipefs type rpc_pipefs (rw,relatime)
zeno:/ on /net/zeno type nfs (rw,nosuid,nodev,relatime,vers=3,rsize=524288,wsize=524288,namlen=255,hard,proto=tcp,timeo=600,retrans=2,sec=sys,mountaddr=192.168.x.y,mountvers=3,mountport=20048,mountproto=udp,local_lock=none,addr=192.168.x.y)
# date
Sat Dec  1 11:40:12 MST 2012
# mount | egrep 'autofs|nfs'
systemd-1 on /proc/sys/fs/binfmt_misc type autofs (rw,relatime,fd=28,pgrp=1,timeout=300,minproto=5,maxproto=5,direct)
sunrpc on /var/lib/nfs/rpc_pipefs type rpc_pipefs (rw,relatime)
-hosts on /net type autofs (rw,relatime,fd=6,pgrp=1060,timeout=60,minproto=5,maxproto=5,indirect)

So in something like in approximately two minutes all NFS mounts are gone.  That is a behaviour which was present there for many years and is rather expected.

My previous contention elsewhere that autofs mounts simply do not expire was not correct.  I was simply not patient enough but timeouts observed on F17 are really trying patience.
Comment 1 Michal Jaegermann 2012-12-01 21:41:47 UTC 
I wonder if bug 831275 has anything in common with this one.  Only 831275 was about Fedora 16 and so far I did not see anything of that sort there.  I do not have around any cifs shares though.
Comment 2 Ian Kent 2012-12-03 02:27:36 UTC 
(In reply to comment #0)
> Description of problem:
> 
> This is the scenario: autofs is running on Fedora 17 client running
> 3.6.7-4.fc17.x86_64 kernel patched with a patch from attachment 654705 [details]
> [details], bug 874372, in order for NFS mounts to work. /etc/auto.master has
> the following in it:
> 
>         /net	-hosts -timeout=60

The problem appears to be the timeout pseudo option, when
used in the master map it needs a "--" not a "-".

The really studid thing is that the timeout option is not
included in the man page where other option, such as the
negative-timeout are and the specification clearly shows
the "--".

Ian
Comment 3 Ian Kent 2012-12-03 02:46:30 UTC 
Created attachment 656333 [details]
Patch - add timeout option description to man page
Comment 4 Michal Jaegermann 2012-12-03 05:56:12 UTC 
(In reply to comment #2)

> 
> The problem appears to be the timeout pseudo option, when
> used in the master map it needs a "--" not a "-".

You are right.  "Copy-and-waste".  I do have '--timeout=60' on f16 but only one "-" on f17.  After correcting that typo the issue disappears.  Where is this paper bag?
Comment 5 Ian Kent 2012-12-03 07:34:22 UTC 
(In reply to comment #4)
> (In reply to comment #2)
> 
> > 
> > The problem appears to be the timeout pseudo option, when
> > used in the master map it needs a "--" not a "-".
> 
> You are right.  "Copy-and-waste".  I do have '--timeout=60' on f16 but only
> one "-" on f17.  After correcting that typo the issue disappears.  Where is
> this paper bag?

Right, but I might as well update the man page while were here.
Comment 6 Fedora Update System 2012-12-03 08:58:06 UTC 
autofs-5.0.6-23.fc17 has been submitted as an update for Fedora 17.
https://admin.fedoraproject.org/updates/autofs-5.0.6-23.fc17
Comment 7 Michal Jaegermann 2012-12-03 19:29:39 UTC 
(In reply to comment #5)
> 
> Right, but I might as well update the man page while were here.

As a matter of fact I am using this '--timeout=...' configuration parameter for a really long time.  Where I got that originally I forgot.  Maybe some README, another form of a documentation, or maybe I even peeked into sources?  I agree that the man page should be explicit about it but chances are that this would not save me from my mistake. :-)
Comment 8 Fedora Update System 2012-12-04 04:54:21 UTC 
Package autofs-5.0.6-23.fc17:
* should fix your issue,
* was pushed to the Fedora 17 testing repository,
* should be available at your local mirror within two days.
Update it with:
# su -c 'yum update --enablerepo=updates-testing autofs-5.0.6-23.fc17'
as soon as you are able to.
Please go to the following url:
https://admin.fedoraproject.org/updates/FEDORA-2012-19632/autofs-5.0.6-23.fc17
then log in and leave karma (feedback).
Comment 9 Fedora Update System 2012-12-12 04:30:47 UTC 
autofs-5.0.6-23.fc17 has been pushed to the Fedora 17 stable repository.  If problems still persist, please make note of it in this bug report.