Bug 458475

Summary: Memory leaks when runing Puppet 0.24.5
Product: [Fedora] Fedora Reporter: Brenton Leanhardt <bleanhar>
Component: rubyAssignee: Jeroen van Meeuwen <vanmeeuwen+fedora>
Status: CLOSED CURRENTRELEASE QA Contact: Fedora Extras Quality Assurance <extras-qa>
Severity: high Docs Contact:
Priority: medium    
Version: 9   
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: 2009-01-31 21:55:02 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:
Attachments:
Description Flags
patch discuessed on http://redmine.ruby-lang.org/issues/show/216 none

Description Brenton Leanhardt 2008-08-08 18:00:19 UTC 
Description of problem:
Memory leaks when running Puppet 0.24.5 with Ruby 1.8.6 patchlevel 230 on F8/F9.

As seen in http://reductivelabs.com/redmine/issues/show/1480, the packages that reproduce the issue on Fedora are :

puppet-server-0.24.5-1.fc9
puppet-0.24.5-1.fc9.noarch0.24.5
ruby-1.8.6.230-4.fc9.i386

While trying to debug I decided to install all the related libraries and tools from scratch. My first test was to install ruby 1.8.7, Puppet 0.24.5 stable from source.  This configuration showed the ruby interpreter Heap staying constantly around 35 MB.  With the F9 package I baselines around 100 MB and quickly reaches over 1 GB using the same test as mentioned in the bug opened on the Reductive Labs site.

http://redmine.ruby-lang.org/issues/show/216 shows information regarding patchlevel 230.  I understand there are good reasons for not using 1.8.7 in Fedora.  I'll try out Matz's patch to see if it solves the Puppetmaster problem and report my findings.
Comment 1 Brenton Leanhardt 2008-08-08 18:48:50 UTC 
Created attachment 313832 [details]
patch discuessed on http://redmine.ruby-lang.org/issues/show/216

I applied this patch to the ruby-1.8.6.230 and rebuilt the patch.  Unfortunately, the memory leak issue with Puppet 0.24.5 is still present.
Comment 2 Brenton Leanhardt 2008-08-08 20:54:33 UTC 
(In reply to comment #1)
> Created an attachment (id=313832) [details]
> patch discuessed on http://redmine.ruby-lang.org/issues/show/216
> 
> I applied this patch to the ruby-1.8.6.230 and rebuilt the patch. 
> Unfortunately, the memory leak issue with Puppet 0.24.5 is still present.

I realized when I updated the specfile I didn't actually apply the patch.  I'll test this again later.
Comment 3 Brenton Leanhardt 2008-08-08 21:04:46 UTC 
(In reply to comment #2)
> (In reply to comment #1)
> > Created an attachment (id=313832) [details] [details]
> > patch discuessed on http://redmine.ruby-lang.org/issues/show/216
> > 
> > I applied this patch to the ruby-1.8.6.230 and rebuilt the patch. 
> > Unfortunately, the memory leak issue with Puppet 0.24.5 is still present.
> 
> I realized when I updated the specfile I didn't actually apply the patch.  I'll
> test this again later.

At first sight the patch seems to be working great.  I'll test it more thoroughly  next week.  I spent most of my afternoon testing with patchlevel 286 which also works very well with Puppet.

Are there any plans to get 286 packaging for Fedora?
Comment 4 Brenton Leanhardt 2008-08-11 21:11:37 UTC 
> Are there any plans to get 286 packaging for Fedora?

Apparently 286 didn't fix all the known security issues: http://www.ruby-lang.org/en/news/2008/08/11/ruby-1-8-7-p72-and-1-8-6-p287-released/
Comment 5 Akira TAGOH 2008-08-12 11:21:38 UTC 
Does -p287 fix this issue?
Comment 6 Brenton Leanhardt 2008-08-14 20:01:05 UTC 
(In reply to comment #5)
> Does -p287 fix this issue?

p287 does fix the problem with p230 originally mentioned in this bug.
Comment 7 Akira TAGOH 2008-08-23 09:32:29 UTC 
Built.
Comment 8 Fedora Admin XMLRPC Client 2009-01-30 12:45:19 UTC 
This package has changed ownership in the Fedora Package Database.  Reassigning to the new owner of this component.
Comment 9 Jeroen van Meeuwen 2009-01-31 21:55:02 UTC 
And released, closing this bug.