Bug 305831 - excessive cpu usage in recovery code
Summary: excessive cpu usage in recovery code
Keywords:
Status: CLOSED INSUFFICIENT_DATA
Alias: None
Product: Red Hat Enterprise Linux 5
Classification: Red Hat
Component: kernel
Version: 5.1
Hardware: All
OS: Linux
low
low
Target Milestone: ---
: ---
Assignee: David Teigland
QA Contact: Red Hat Kernel QE team
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2007-09-25 18:51 UTC by David Teigland
Modified: 2009-09-03 16:51 UTC (History)
3 users (show)

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Clone Of:
Environment:
Last Closed: 2009-05-01 19:13:47 UTC
Target Upstream Version:
Embargoed:


Attachments (Terms of Use)

Description David Teigland 2007-09-25 18:51:04 UTC
Description of problem:

The dlm recovery code calls schedule() in a lot of lock loops
to prevent:

1) the softlock watchdog from going off
2) openais cluster membership messages for being delayed past
   the configured timeout

We want to investigate:

- will cond_resched() work as well, and more efficiently than schedule()?
  (I expect so)
- exactly what loops are taking so long (watchdog is 10 sec) and why?
  are there really that many locks and/or are we doing that much work
  on each one that it can take 10 sec?
- why does this seem to appear on ia64 regularly and other arch's rarely?

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


How reproducible:


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


Expected results:


Additional info:

Comment 1 David Teigland 2008-06-24 17:25:27 UTC
Dean, do you still get softlockups on ia64?



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