Bug 118352

Summary: (VM) setfiles exhausts available memory
Product: [Fedora] Fedora Reporter: Dan Williams <dcbw>
Component: kernelAssignee: Dave Jones <davej>
Status: CLOSED WORKSFORME QA Contact: Brian Brock <bbrock>
Severity: medium Docs Contact:
Priority: medium    
Version: 2CC: jmorris, pfrields, pgraner, sds
Target Milestone: ---   
Target Release: ---   
Hardware: All   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2004-12-09 18:23:30 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: ---
Bug Depends On:    
Bug Blocks: 114963, 123268, 136451    
Attachments:
Description Flags
kernel messages when running out of memory running setfiles none

Description Dan Williams 2004-03-15 21:03:56 UTC
Description of problem:
When running 'make relabel' on a drive with large numbers of files
(150000) setfiles eventually consumes all available memory and begins
to swap.  It eventually uses all swap and hangs machine

Version-Release number of selected component (if applicable):
policycoreutils-1.6-8

How reproducible:
every time on my x86_64 box

Steps to Reproduce:
1.
2.
3.
  
Actual results:


Expected results:


Additional info:

Comment 1 Daniel Walsh 2004-03-17 15:35:30 UTC
This looks like it might be a kernel problem.

The app is not growing but free memory is dropping.

> Stephen Smalley says he sees similar behaviour when running
> Ah, good point - I was just looking at memory consumption by the
> process, not at the massive bleeding of overall free memory during a
> setfiles run.  Looks like find / -xdev -ls > /dev/null yields the 
> same result; requires further investigation.

Memory seems to free up when the app exits.




Comment 2 Dan Williams 2004-03-17 16:29:30 UTC
Created attachment 98608 [details]
kernel messages when running out of memory running setfiles

This is using kernel 2.6.3-2.1.253 on x86-64 box

Comment 3 James Morris 2004-04-08 16:26:08 UTC
I've looked into this on i386 and have not observed anything unusual
(e.g. memory pressure reclaims dcache, ext3 inode cache slab entries
just fine).

Can you please test with the current Fedora kernel?

If you still see the problem, are you able to duplicate it on an i386 box?


Comment 4 Alan Cox 2004-05-03 16:35:02 UTC
This looks a lot like the updatedb problem report. You might want to
try Arjan's latest and greatest


Comment 5 Dave Jones 2004-12-07 06:15:55 UTC
still a problem with the 2.6.9 updates ?


Comment 6 Dan Williams 2004-12-09 18:23:30 UTC
Haven't experienced for some time, if it occurs again I'll reopen or
file a new one.