Bug 1396403

Summary: Trying to access a non-existent directory using automount results in 4 minute hang as not checking the local mount availability
Product: Red Hat Enterprise Linux 6 Reporter: Yongcheng Yang <yoyang>
Component: autofsAssignee: Ian Kent <ikent>
Status: CLOSED ERRATA QA Contact: Yongcheng Yang <yoyang>
Severity: medium Docs Contact: Milan Navratil <mnavrati>
Priority: high    
Version: 6.9CC: ajmitchell, bcodding, bfields, chunwang, eguan, fs-qe, ikent, mnavrati, steved, swhiteho, tlavigne
Target Milestone: rcKeywords: Regression
Target Release: ---   
Hardware: All   
OS: Linux   
Whiteboard:
Fixed In Version: autofs-5.0.5-131.el6 Doc Type: Bug Fix
Doc Text:
Cause: A server availability probe is not done for mount requests that autofs thinks are local because a bind mount on the local machine should be able to be used. If the bind mount fails an NFS mount on the local machine then tried. Consequence: But if the NFS server is not running on the local machine the mount attempt can suffer a lengthy time out before failing. Fix: To resolve this an availability probe has been added to the case where a bind mount is first tried but fails and autofs falls back to trying to use an NFS server on the local machine. Result: If a bind mount on the local machine fails the fall back to trying an NFS mount on the local machine fails quickly if the local NFS server is not running.
 Story Points: ---
Clone Of:
: 1420574 (view as bug list) Environment:
Last Closed: 2017-03-21 11:47:11 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:
Bug Depends On:    
Bug Blocks: 1401884, 1420574    
Attachments:
Description Flags
Patch - fix inconsistent signed usage for __rpc_ping()
none
Patch - check NFS server availability on local mount fallback none

Comment 13 Steve Whitehouse 2017-01-05 15:15:24 UTC 
We need to figure out a solution here fairly soon, since time is limited to make changes in 6.9. Is this really a regression?

Steve D / Ian what is the lowest risk option that is open to us?
Comment 14 Ian Kent 2017-01-06 00:45:05 UTC 
(In reply to Steve Whitehouse from comment #13)
> We need to figure out a solution here fairly soon, since time is limited to
> make changes in 6.9. Is this really a regression?
> 
> Steve D / Ian what is the lowest risk option that is open to us?

I think mount.nfs(8) needs work but I also can't leave autofs
open to this.

I should be able to add RPC ping check for this special case
alone without much effort and since the mount ultimately fails
it shouldn't introduce a regression.

I'll do that.
Comment 15 Ian Kent 2017-01-06 00:46:45 UTC 
Steve hasn't devel_acked this so I guess it will be ok for me to
take this for autofs, we can always change it back if need be.
Comment 24 Ian Kent 2017-01-10 01:57:21 UTC 
Created attachment 1238963 [details]
Patch - fix inconsistent signed usage for __rpc_ping()

Missing autofs-5.0.7 patch for rpc_ping.
Comment 25 Ian Kent 2017-01-10 01:58:41 UTC 
Created attachment 1238964 [details]
Patch - check NFS server availability on local mount fallback
Comment 31 Yongcheng Yang 2017-01-11 02:56:42 UTC 
Moving to VERIFIED according to test logs of Comment #30.
Comment 32 Yongcheng Yang 2017-02-09 03:37:32 UTC 
Updating the bug description according to Comment #12 and Comment #14.

Thanks Ian for the explanation.
Comment 36 errata-xmlrpc 2017-03-21 11:47:11 UTC 
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://rhn.redhat.com/errata/RHBA-2017-0780.html