pcp
[Top] [All Lists]

Re: [pcp] 3.7.0 QA results

To: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Subject: Re: [pcp] 3.7.0 QA results
From: Nathan Scott <nathans@xxxxxxxxxx>
Date: Wed, 6 Mar 2013 17:40:39 -0500 (EST)
Cc: pcp@xxxxxxxxxxx
Delivered-to: pcp@xxxxxxxxxxx
In-reply-to: <5137B9BB.701@xxxxxxxxxxxxxxxx>
Reply-to: Nathan Scott <nathans@xxxxxxxxxx>
Hi Ken,

("pcp && pmconfig -L" output would help to track down some of these)

----- Original Message -----
> 
> No analysis or attempt to repeat here, just reporting failures for
> first 5 QA hosts.
> vm00 3.7.0 x86_64 Ubuntu 12.04 (precise)
> 
> 
> 169 different error message for IPC failure
> 197 next fd is 4 not 9
> 430 repeated instances of extra pmlogger being started

I've hunted down 430 recently and it has proved elusive - seems to
be some indeterminism in the test, as earlier in the week I saw it
a few times but yesterday could not reproduce it at all.

> vm01 3.7.0 i686 Ubuntu 12.10 (quantal)
> 
> 
> 197 next fd is 4 not 9
> 372 Perl Spreadsheet::ReadSXC module not installed
> 430 repeated instances of extra pmlogger being started
> 532 bad QA test failure /tmp/21467.out.1: No such file or directory
> vm02 3.7.0 i686 openSUSE 12.1
> 
> 
> 093 unexpected number of reads for foo archive
> 388 valgrind errors Conditional jump or move depends on uninitialised
> value(s) @ interp.c:1435
> 393 valgrind errors Conditional jump or move depends on uninitialised
> value(s) @ interp.c:1435
> 394 valgrind errors Conditional jump or move depends on uninitialised
> value(s) @ interp.c:1435
> 430 repeated instances of extra pmlogger being started
> 459 valgrind errors Conditional jump or move depends on uninitialised
> value(s) @ interp.c:1233
> 511 sar2pcp failed ... mismatched tag at line 75, column 5, byte 4773
> at /usr/lib/perl5/site_perl/5.14.2/XML/TokeParser.pm

(broken version of sadf? - like opensuse has/had?  not sure we can
fix this, as its invalid XML being generated by sysstat tools).

> vm03 3.7.0 x86_64 Fedora 18 (Spherical Cow)
> 
> 
> 
> 169 pmcd SIGHUP didn't restart the dumb_pmda agent
> 244 different error message for IPC failure, Error: phony fake_cisco
> agent died prematurely, multiple PMDA/PMCD permission failures, ...
> 255 broken_2_0 PMDA died prematurely
> 279 pmcd failed to restart after reconfig
> 420 Updating /etc/hosts failed
> 555 Botch: using rsyslog, but I don't know how to stop/start it

Could be systemd fallout, its likely f18 rsyslog got systemd start/
stop service treatment.

> 652 Extra pmlogger start, different metrics from systemd PMDA
> vm04 3.7.0 i586 CentOS 5.9 (Final)
> 
> 087 different number of log reads
> 193 missing data from archive
> 273 pmcd PMDA died?

?  (its a dso PMDA though)

> 316 floating point arithmetic non-determinism
> 393 valgrind memory leak 11,274 bytes in 1,288 blocks are possibly
> lost @ pmns.c:1079
> 430 repeated instances of extra pmlogger being started
> 459 valgrind memory leak 88,920 bytes in 1,170 blocks are possibly
> lost @ interp.c:688
> 482 wrong answers for arithmetic!
> 707 Not OK
> 708 Not OK
> 709 pmcollectl failed
> 710 check_import python script gave non-zero result

Last four look like total python module failure, probably some
packaging/install failure there.


... and I was feeling so good before about having zero failures on
my couple of hosts for the first time in a long time.  :)

cheers.

--
Nathan

<Prev in Thread] Current Thread [Next in Thread>