Bug 181646

Summary:	executables shared over NFS crash with SIGBUS when new executables are released.
Product:	Red Hat Enterprise Linux 3	Reporter:	Sev Binello <sev>
Component:	nfs-utils	Assignee:	Steve Dickson <steved>
Status:	CLOSED NOTABUG	QA Contact:	Ben Levenson <benl>
Severity:	high	Docs Contact:
Priority:	medium
Version:	3.0
Target Milestone:	---
Target Release:	---
Hardware:	i386
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2006-02-21 19:30:59 UTC	Type:	---
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:

Description Sev Binello 2006-02-15 18:17:55 UTC

From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.12) Gecko/20050921 Red Hat/1.7.12-1.1.3.2

Description of problem:
We have had a consistent problem of executables shared over NFS crashing with a SIGBUS when new executables are released.

   Before a new executable replaces an older one with a cp,
   we mv the executable to a fallback dir and rename it.
   We have 2 renaming schemes,
   either .old, .older, .oldest or  append a date to the old executable.
     This used to work fine when the nfs sever was solaris,
   however since we migrated it to redhat linux WS,
   we find that the executables crash with a SIGBUS.
     The executables don't necessarily crash right away,
   as a matter of fact we can't intentionally reproduce the crash.
   Which leads me to believe it is cache related.

   Has anyone experienced a similar problem,
    and come up with a solution ?


Version-Release number of selected component (if applicable):
nfs-utils-1.0.6-42EL 

How reproducible:
Couldn't Reproduce

Steps to Reproduce:
1
2.
3.
  

Additional info:

I cant intentionally reproduce it,
but it does happen often.
Usually users will leave applications up and idle over an extended period
and when they go back to reuse it, the application will crash
with a SIGBUS. The problem is not restricted to any one application.

We are running Red Hat Enterprise Linux WS release 3 (Taroon Update 6)
 2.4.21-37.ELsmp

Comment 1 Sev Binello 2006-02-21 19:30:59 UTC

This was not a bug. 
Here's some mail from Trond Mykelbust explaining the problem ...

    It isn't. It is a consequence of using the (admittedly _very_ stupid)
     default server setting of "subtree_check", and then renaming into a
     different directory (man 5 exports).

    You have one of 2 possible solutions:

    1) Edit your /etc/exports, and add no_subtree_check options to the
       entries for your exported filesystem.
    OR
    2) rename the executables into the same directory.