Some changes already in oss tree here, but pushes were delayed pending last
release.
Changes committed to git://git.pcp.io/kenj/pcp dev
Ken McDonell (19):
libpcp/access.c: remove extra PM_UNLOCK
libpcp/loop.c: long dead code ... remove it
qa/.gitignore: 926.out is not checked in
qa/518: generalize previous fix to deal with line breaks in pcp(1) output
qa/832: better handling for warnings from SASL code
qa/admin/check-vm: yet another package providing the RRDs perl module
qa/832: rescind one of the last filter changes
qa/720: some more fiddling about with the filtering
qa/660: strip more non-deterministic output
qa/947: accept 3, 4 or 5 lines of output
qa/common.check: for valgrind add --gen-suppressions=all
qa/valgrind-suppress-3.5.0: add suppressions for qa/720 on vm24
qa/admin/pcp-daily: be more anal about flushing the page cache before
shutdown
libpcp/AF.c: add comment to capture current state of signal safeness (or
not)
qa/145: deal with some more optional pmcd feature bits
qa/179: increase sampling interval to improve reliability
qa/578: add a bit more diagnostics
qa/753: reduce precision of output
qa/115 & 769: chkconfig does not work on Gentoo
Nathan Scott (3):
qa: add missing no-python notrun check to 904
qa: ensure 904 output consistent between python versions
libpcp: fix non-secure-socket builds requesting containers
Frank Ch. Eigler (1):
configure: handle 'check' as a value for $do_papi.
configure | 2
configure.ac | 4
qa/.gitignore | 1
qa/115 | 9
qa/145 | 12
qa/179 | 8
qa/179.out | 4
qa/518 | 2
qa/578 | 3
qa/660 | 4
qa/660.out.4 | 3
qa/660.out.46 | 3
qa/720 | 4
qa/753 | 4
qa/753.out | 8
qa/769 | 4
qa/832 | 22 -
qa/904 | 3
qa/904.out | 6
qa/947 | 7
qa/947.out | 3
qa/admin/check-vm | 6
qa/admin/pcp-daily | 2
qa/common.check | 4
qa/src/test_pmnswalk.python | 8
qa/valgrind-suppress-3.5.0 | 41 ++
src/libpcp/src/AF.c | 20 +
src/libpcp/src/access.c | 1
src/libpcp/src/auxconnect.c | 13
src/libpcp/src/loop.c | 866 --------------------------------------------
30 files changed, 157 insertions(+), 920 deletions(-)
Details ...
commit f31d66236af37a83be33fa97774c6b736aacf841
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Thu Mar 5 20:12:31 2015 +1100
qa/115 & 769: chkconfig does not work on Gentoo
We don't know how to make chkconfig work for Gentoo, so accept
this and adjust QA expectations accordingly.
commit 0c56004ff120ff03c1f4bddae99bd059f1845b29
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Thu Mar 5 20:08:25 2015 +1100
qa/753: reduce precision of output
Use pmval -f 0 rather than -f 1 so numbers are reported as
6, rather than 6.0 ... which was failing on some slower hosts
when the value was reported as 5.8 or 5.9.
commit 67fd032d188487ae8580c5a1701d5c5f2dd431d9
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Thu Mar 5 16:48:48 2015 +1100
qa/578: add a bit more diagnostics
May help diagnose intermittent failures.
commit cd8008ff61973245d8952ef5a57d6352ae80f873
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Thu Mar 5 08:49:53 2015 +1100
qa/179: increase sampling interval to improve reliability
Logging with 20msec intervals was not deterministic on some slower
platforms, and the interval is not relevant for the function the
test is performing.
commit 639cb218cdff2b950ea2fd11b0d8549ff2c8c116
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Thu Mar 5 07:26:50 2015 +1100
qa/145: deal with some more optional pmcd feature bits
Was failing on grundy and vm05 with no authentication and no
secure sockets.
Rewriting the PDU dump output is getting really messy here!
commit 9f1a7ad36b201ea876994c908f15045fc6f412b3
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Wed Mar 4 10:47:17 2015 +1100
libpcp/AF.c: add comment to capture current state of signal safeness (or
not)
commit 3bbe2b06d21b30e5d7133add4e96ac5f807b8b7d
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Wed Mar 4 10:43:56 2015 +1100
qa/admin/pcp-daily: be more anal about flushing the page cache before
shutdown
I was seeing some filesystem curruption after VM shutdown, and this
is a valiant (and old fashioned) way to try guard against this.
commit 9bba3bf1e19fb45955b8a0b1af29216311339d63
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Wed Mar 4 10:43:09 2015 +1100
qa/valgrind-suppress-3.5.0: add suppressions for qa/720 on vm24
commit 2a07b179e95b7d3e648c1c7b7a53a2597ada6044
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Wed Mar 4 10:41:07 2015 +1100
qa/common.check: for valgrind add --gen-suppressions=all
When valgrind finds a problem may as well get the potential
suppression in the .out.bad file in case it is not a PCP
problem.
commit c68b7aea47fa9b76b5cbd7ad9140cf8900b4149a
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Tue Mar 3 09:27:33 2015 +1100
qa/947: accept 3, 4 or 5 lines of output
There is non-determinism between the sleep+kill in the script
and the pmlogger timing, so expecting exactly 4 lines of output
is a bit brittle.
commit 2a7512d2ee1a1faef317e56ee66fd5bf4df01b6f
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Tue Mar 3 09:26:38 2015 +1100
qa/660: strip more non-deterministic output
commit 921c40c9f56cf5403176573e8964b054a4227374
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Tue Mar 3 09:26:03 2015 +1100
qa/720: some more fiddling about with the filtering
commit cf2b54c1ab7bc9741c4fda07bc219a66244c6573
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Tue Mar 3 07:48:12 2015 +1100
qa/832: rescind one of the last filter changes
send-mail: Cannot open mail:25
is not a SASL issue, it is a FC19 et al smtp mail configuration
breakage issue.
Don't filter here, needs to be fixed on platform setup.
commit 2622327458907319367ae98c4048c01fccc4dfe0
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Tue Mar 3 06:38:19 2015 +1100
qa/admin/check-vm: yet another package providing the RRDs perl module
commit 4bf55dcc1e396f0fe6498d8203a3e8b540fcb5a3
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Tue Mar 3 06:36:20 2015 +1100
qa/832: better handling for warnings from SASL code
"Alert: Parameter Error" can appear in multiple contexts and
from multiple places in the script. The previous filtering was
not capturing and removing all of 'em.
commit 37ff5e02ef1654a4399febde1237be36587a235a
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Tue Mar 3 06:27:16 2015 +1100
qa/518: generalize previous fix to deal with line breaks in pcp(1) output
commit 741b05ab3ecb794495245ede65ce484de9ea657f
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Tue Mar 3 06:26:30 2015 +1100
qa/.gitignore: 926.out is not checked in
commit 6b2fb91fc15bd64529d181ac0eb5ac50a5ca6d2f
Author: Frank Ch. Eigler <fche@xxxxxxxxxx>
Date: Mon Mar 2 10:30:09 2015 -0500
configure: handle 'check' as a value for $do_papi.
commit 37c273044c457be4f88c22becc1364f9ac19945d
Author: Nathan Scott <nathans@xxxxxxxxxx>
Date: Mon Mar 2 20:34:34 2015 +1100
libpcp: fix non-secure-socket builds requesting containers
commit be13d9fb98b49cbdc96985634f4a81e218fc9300
Author: Nathan Scott <nathans@xxxxxxxxxx>
Date: Mon Mar 2 18:37:45 2015 +1100
qa: ensure 904 output consistent between python versions
commit e53ea161c31a54b69e9cb2cae5a4bb5aab5bb2df
Author: Nathan Scott <nathans@xxxxxxxxxx>
Date: Mon Mar 2 18:32:46 2015 +1100
qa: add missing no-python notrun check to 904
commit 9e2e4bb65addb795351b27320a52a07400e111d0
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Mon Mar 2 07:56:15 2015 +1100
libpcp/loop.c: long dead code ... remove it
commit a9fe35eb2b1ade90520eb1f772a38a07a995f227
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Mon Mar 2 07:42:09 2015 +1100
libpcp/access.c: remove extra PM_UNLOCK
On "no memory" error path, so most unlikely that this has been seen
in practices.
Partial fix for issues raised in
http://oss.sgi.com/bugzilla/show_bug.cgi?id=1055
|