Bug 179086 - fork() slow when shm attached to process
fork() slow when shm attached to process
Status: CLOSED DUPLICATE of bug 185386
Product: Red Hat Enterprise Linux 4
Classification: Red Hat
Component: kernel (Show other bugs)
4.0
All Linux
medium Severity medium
: ---
: ---
Assigned To: Larry Woodman
Brian Brock
:
Depends On:
Blocks: 170416
  Show dependency treegraph
 
Reported: 2006-01-27 04:45 EST by Bastien Nocera
Modified: 2007-11-30 17:07 EST (History)
2 users (show)

See Also:
Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of:
Environment:
Last Closed: 2006-04-19 15:22:24 EDT
Type: ---
Regression: ---
Mount Type: ---
Documentation: ---
CRM:
Verified Versions:
Category: ---
oVirt Team: ---
RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: ---


Attachments (Terms of Use)

  None (edit)
Description Bastien Nocera 2006-01-27 04:45:25 EST
kernel-2.6.9-22.0.1.EL

The fork() call time grows linearly based on amount of SysV memory attached to
the process.

Fix available upstream:
http://www.kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=d992895ba2b27cf5adf1ba0ad6d27662adc54c5e

Upstream discussion:
http://groups.google.com/group/linux.kernel/browse_thread/thread/937a999061644f1d/be39ede8f991e847
Comment 1 Chris Snook 2006-03-02 18:23:55 EST
Support has received calls from customers of Ontario Systems, who recommend
running kernels with the lazy fork patch.  For apps that use large shared memory
spaces and fork to service each request, the performance impact can be orders of
magnitude.

The downside is that it's possible to have an OOM condition when the app
accesses a new part of the shared segment and the kernel needs to copy a PTE, a
very long time after the fork itself, thus hiding the cause.  If we implement
this, we should at least add an appropriate error message in oom-killer as well.
Comment 2 Jason Baron 2006-04-19 15:22:24 EDT

*** This bug has been marked as a duplicate of 185386 ***

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