Skip to contentSkip to conversation

user reporting memory leak with ldap maps

DISCUSSION Guillaume Rousse 2 participants 7 messages Apr 14, 2008 – Jun 13, 2008 Kernel automounter

lore
Share & tools
lore
#0 lore raw

A mdv user complains about a memory leak in autofs 5.0.2. I can't reproduce myself, but I don't have the same version available for testing.

See https://qa.mandriva.com/show_bug.cgi?id=39911 for full report.

Signature · 3 lines
Guillaume Rousse
Moyens Informatiques - INRIA Futurs
Tel: 01 69 35 69 62
maintainer
#1 lore raw
Quoted 4 lines · Guillaume Rousse

A mdv user complains about a memory leak in autofs 5.0.2. I can't reproduce myself, but I don't have the same version available for testing.

See https://qa.mandriva.com/show_bug.cgi?id=39911 for full report.

That's quite possible but why hasn't any pertinent information been obtained from the reporter?

What are the maps in use?
What is the memory usage at startup, after 2 hours, 4 hours, at least two samples the following day, etc. (for at least two days)? Does the same thing happen if they setup file maps for everything instead?

What process did you follow when trying to reproduce it?

As far as the version goes I don't know what may have been added to make up that package.

Ian

maintainer
#2 lore raw
Quoted 1 line · Ian Kent

What are the maps in use?

I see we now have the maps used.

Ian

#3 lore raw
In reply to Ian Kent
Quoted 6 lines · Ian Kent
Quoted 4 lines · Guillaume Rousse

A mdv user complains about a memory leak in autofs 5.0.2. I can't reproduce myself, but I don't have the same version available for testing.

See https://qa.mandriva.com/show_bug.cgi?id=39911 for full report.

Quoted 8 lines

That's quite possible but why hasn't any pertinent information been obtained from the reporter?

What are the maps in use?
What is the memory usage at startup, after 2 hours, 4 hours, at least two samples the following day, etc. (for at least two days)? Does the same thing happen if they setup file maps for everything instead?

I did ask.

Quoted 4 lines

What process did you follow when trying to reproduce it?

As far as the version goes I don't know what may have been added to make up that package.

If you're refering to mandriva specific modifications, everything is available at
http://svn.mandriva.com/cgi-bin/viewvc.cgi/packages/cooker/autofs/releases/5.0.2/8mdv2008.1/

Most patches are upstream patches, excepted the following ones:

Patch101:       autofs-5.0.2-set-default-browse-mode.patch
Patch102:       autofs-5.0.2-separate-config-files.patch
Patch103:       autofs-5.0.2-rename-configuration-file.patch
Patch105:       autofs-5.0.1-rc3-comment-default-master-map.patch
Signature · 3 lines
Guillaume Rousse
Moyens Informatiques - INRIA Futurs
Tel: 01 69 35 69 62
maintainer
#4 lore raw
Quoted 17 lines · Guillaume Rousse

Ian Kent a écrit :

Quoted 6 lines · Guillaume Rousse
Quoted 4 lines

A mdv user complains about a memory leak in autofs 5.0.2. I can't reproduce myself, but I don't have the same version available for testing.

See https://qa.mandriva.com/show_bug.cgi?id=39911 for full report.

Quoted 8 lines

That's quite possible but why hasn't any pertinent information been obtained from the reporter?

What are the maps in use?
What is the memory usage at startup, after 2 hours, 4 hours, at least two samples the following day, etc. (for at least two days)? Does the same thing happen if they setup file maps for everything instead?

I did ask.

Quoted 7 lines
Quoted 4 lines

What process did you follow when trying to reproduce it?

As far as the version goes I don't know what may have been added to make up that package.

If you're refering to mandriva specific modifications, everything is available at
http://svn.mandriva.com/cgi-bin/viewvc.cgi/packages/cooker/autofs/releases/5.0.2/8mdv2008.1/

That's what I need.
I'll have a go at setting up a test.

Quoted 6 lines

Most patches are upstream patches, excepted the following ones:

Patch101:       autofs-5.0.2-set-default-browse-mode.patch
Patch102:       autofs-5.0.2-separate-config-files.patch
Patch103:       autofs-5.0.2-rename-configuration-file.patch
Patch105:       autofs-5.0.1-rc3-comment-default-master-map.patch

_______________________________________________
autofs mailing list
[email protected]
http://linux.kernel.org/mailman/listinfo/autofs

#5 lore raw
In reply to Guillaume Rousse
Quoted 9 lines · Guillaume Rousse
Quoted 8 lines · Ian Kent
Quoted 4 lines · Guillaume Rousse

A mdv user complains about a memory leak in autofs 5.0.2. I can't reproduce myself, but I don't have the same version available for testing.

See https://qa.mandriva.com/show_bug.cgi?id=39911 for full report.

That's quite possible but why hasn't any pertinent information been obtained from the reporter?

Quoted 6 lines
Quoted 5 lines

What are the maps in use?
What is the memory usage at startup, after 2 hours, 4 hours, at least two samples the following day, etc. (for at least two days)? Does the same thing happen if they setup file maps for everything instead?

I did ask.

The user just updated the ticket with memory usage for both file and ldap-based configuration.

Signature · 3 lines
Guillaume Rousse
Moyens Informatiques - INRIA Futurs
Tel: 01 69 35 69 62
maintainer
#6 lore raw
Quoted 10 lines · Guillaume Rousse
Quoted 9 lines · Guillaume Rousse
Quoted 8 lines · Ian Kent
Quoted 4 lines · Guillaume Rousse

A mdv user complains about a memory leak in autofs 5.0.2. I can't reproduce myself, but I don't have the same version available for testing.

See https://qa.mandriva.com/show_bug.cgi?id=39911 for full report.

That's quite possible but why hasn't any pertinent information been obtained from the reporter?

Quoted 8 lines
Quoted 6 lines
Quoted 5 lines

What are the maps in use?
What is the memory usage at startup, after 2 hours, 4 hours, at least two samples the following day, etc. (for at least two days)? Does the same thing happen if they setup file maps for everything instead?

I did ask.

The user just updated the ticket with memory usage for both file and ldap-based configuration.

Thank them for this information please.
I'll test this out when I've got on top of the current issues I'm working on.

I have to mention that I spent a fair amount of time checking memory usage during v5 development, but it's been a while since I've done this now and there have been quite a few changes since. The quick check I did after your post indicated (as I expected) that this most likely isn't just a couple of simple autofs errors as the valgrind report suggested that functions outside of autofs were not freeing memory, which may or may not really be the case or the report could be completely misguided.

Ian

_______________________________________________
autofs mailing list
[email protected]
http://linux.kernel.org/mailman/listinfo/autofs