Bug 80353 - Probable memory leaks in anaconda
Summary: Probable memory leaks in anaconda
Alias: None
Product: Red Hat Public Beta
Classification: Retired
Component: anaconda
Version: phoebe
Hardware: i386
OS: Linux
Target Milestone: ---
Assignee: Michael Fulbright
QA Contact: Mike McLean
Depends On:
TreeView+ depends on / blocked
Reported: 2002-12-24 22:11 UTC by jfm2
Modified: 2007-04-18 16:49 UTC (History)
1 user (show)

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Clone Of:
Last Closed: 2003-01-24 22:41:20 UTC

Attachments (Terms of Use)

Description jfm2 2002-12-24 22:11:17 UTC
From Bugzilla Helper:
User-Agent: Mozilla/5.0 Galeon/1.2.6 (X11; Linux i586; U;) Gecko/20020830

Description of problem:
By the end of installation anaconda was using 250 megs of main memory and 170
megs of swap for a total of 420 megs.  My guess is that it does
not free the memory for the photos.  There are many of them and they
change frequently.  They are very nice but if you don't find where anaconda is
leaking you should consider reducing their number and frequency of reloading.

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

How reproducible:

Steps to Reproduce:
1.Install in graphic mode
2. Run free from time to time during the installation

Additional info:

Comment 1 Jeremy Katz 2003-01-02 21:16:37 UTC
There were some leaks in the version of pygtk2 used in phoebe.  1.99.14 fixes these

Comment 2 Mike McLean 2003-01-14 01:01:42 UTC
this seems to be fixed

Comment 3 jfm2 2003-01-23 07:14:13 UTC
In Phoebe2 by the end of installation it was using some 350 megs of memory+swap
(I installed less things than when I opened the bug).
Checks during the installation showed constant increase during package
installation phase

Comment 4 Mike McLean 2003-01-24 22:41:20 UTC
The output of free (or contents of /proc/meminfo) can be misleading.  How the
kernel manages physical resources like memory and swap space is separate from
how anaconda (or any other userland process) manages the memory allocated to it
by the kernel. 

The best indication of anaconda's memory use is /proc/$PID/status, where $PID is
the pid of anaconda.  I have run some additional tests and anaconda does not
appear to be leaking memory in Phoebe public beta 2.

Usually when anaconda does have a memory leak (and I think there was a minor one
in Phoebe public beta 1), I will see a few of my testcases stall out completely,
particularly some of the low-memory testcases.  I did not see this with the
second public beta.

Re-closing this bug.

Note You need to log in before you can comment on or make changes to this bug.