Bug 476690
| Summary: | HTS-5.3-9 Profiler Test Fails | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Product: | [Retired] Red Hat Hardware Certification Program | Reporter: | Mark Coskey <mark.coskey> | ||||||||
| Component: | Test Suite (tests) | Assignee: | Rob Landry <rlandry> | ||||||||
| Status: | CLOSED ERRATA | QA Contact: | Lawrence Lim <llim> | ||||||||
| Severity: | high | Docs Contact: | |||||||||
| Priority: | low | ||||||||||
| Version: | 5.3 | CC: | averma, gnichols, gregg.shick, quang.duong, rlandry, tools-bugs, tyan, ykun | ||||||||
| Target Milestone: | --- | Keywords: | Documentation, Reopened | ||||||||
| Target Release: | --- | ||||||||||
| Hardware: | All | ||||||||||
| OS: | Linux | ||||||||||
| Whiteboard: | |||||||||||
| Fixed In Version: | Doc Type: | Bug Fix | |||||||||
| Doc Text: |
Cause:
======
The PROFILER test fails for "no sample file found".
Consequence:
============
The PROFILER test could not pass.
Fix:
====
Add the check for Timer configuration.
Result:
=======
The PROFILER test works well now.
|
Story Points: | --- | ||||||||
| Clone Of: | Environment: | ||||||||||
| Last Closed: | 2009-08-27 16:35:42 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
Mark Coskey
2008-12-16 15:51:35 UTC
Created attachment 327117 [details]
HTS results package
Looks like there some bogus output from opcontrol... /usr/bin/opcontrol: line 1118: echo: write error: Invalid argument ...which we'll need to check into to see if it's a bug in hts or oprofile or whatever. In addition please add "options oprofile timer=1" to modprobe.conf and rerun the oprofiler test. This may work around the profile test failure, but it looks like oprofile was trying to use the more exact hardware modes but failed to do so. Created attachment 328877 [details]
failed profiler on amd
Created attachment 328879 [details]
5.3-12 passing output on intel box
Hi Gregg, Can you try adding... options oprofile timer=1 ...to modprobe.conf on the AMD box and rerun the test? You may also need to verify that oprofile isn't already loaded before running the test again to ensure the option takes effect. ...or was that already done in these later runs? Profiler passed on AMD box using this command (was run with hts5.3-12) Quang Is this still a problem or does the timer mode and/or hts-5.3-15 resolve this issue? I'm going to go ahead and close this presuming either the later hts package or the timer mode resolves the issue. Rob 5.3-15 did not resolve this issue. We are still using the workaround. Do you have a later hts to try? Hi Gregg, workaround meaning timer mode or something else? There is an hts-5.3-17 in RHN now, though I don't think there are any profiler test fixes in it. Rob Yep the timer mode change: 1. Open modprobe.conf 2. Add "options oprofile timer=1" We still have to do it with -15. Gregg, The -xen kernel should default to the timer mode, also unrecognised processors by oprofile should default to timer mode. Is the same behavior noted when using the BM kernel and is this on all system or specific one with a specific processor? Rob We saw it mostly on AMD procs, but also some Intels. I don't have a more recent list since we have gone to making sure that line is present but we will start running profiler without any manually edits first to identify the procs. Hey Greg, v7-1.0-6 includes a patch to automatically add the timer=1 option as required which should hopefully avoid the need to manually provide it. From what we've been able to determine there's an issue with oprofile version included in RHEL4 where it and the nmiwatchdog will have a resource collision if the specific hardware does not include dedicated support in oprofiler; specifically where opofiler tries to fall back to the nmitimer which is already claimed by the nmiwatchdog; this seems to be the root cause for the many cases of RHEL4 needing the timer=1 option, as this falls back even farther into a software timer operation. Verified v7-1.0-14.el5 profiler test passed on RHEL5.4 Server x86_64 system. Release note added. If any revisions are required, please set the "requires_release_notes" flag to "?" and edit the "Release Notes" field accordingly. All revisions will be proofread by the Engineering Content Services team. New Contents: Cause: ====== The PROFILER test fails for "no sample file found". Consequence: ============ The PROFILER test could not pass. Fix: ==== Add the check for Timer configuration. Result: ======= The PROFILER test works well now. An advisory has been issued which should help the problem described in this bug report. This report is therefore being closed with a resolution of ERRATA. For more information on therefore solution and/or where to find the updated files, please follow the link below. You may reopen this bug report if the solution does not work for you. http://rhn.redhat.com/errata/RHBA-2009-1234.html The needinfo request[s] on this closed bug have been removed as they have been unresolved for 1000 days |