Changes committed to git://git.pcp.io/pcp.git master
Ken McDonell (9):
qa/519: tweak filtering after recent pmie change
qa/622: updated output after pmie report format change
qa/583: tweak filtering after recent pmie change
qa/294: track recent pmie changes
src/pmie/show.c: small changes afer FQDN changes
pmie man page: minor wordsmithing around (new) %c format item
qa/admin/check-vm: report correct Debian package for JSON Perl module
linux pmda: add Intel copyright to proc_partions.c
qa/869: (new) check NVME disk names for archive from linux pmda
Frank Ch. Eigler (8):
pmmgr: subtarget-containers support
pmmgr: subtarget-containers qa/668
pmmgr: lengthen subtarget-containers qa/668
pmmgr subtarget-containers: disable by default
pmmgr container pmie qa: track expected number of pmie.log files
PR1105: fqdn -h support for pmie
pmie-fqdn qa/983: pmie -h test
pmie-fqdn: support "%c" substitution in action strings
Nathan Scott (4):
pmmgr PR1105: enable pmie for subtarget-containers
pmdaroot: add a metric to aid mapping containers to their cgroups
build: FreeBSD build fix for 10.2 with dtrace probes auto-enabled
pmmgr: remove unhelpful comments about libpcp and containers
Lukas Berk (1):
Modify qa/812 to use hostzone options (fixes buildbot timezone based
failures)
Suruchi Kadu (1):
linux pmda: add support for NVME devices
man/man1/pmie.1 | 27
man/man1/pmmgr.1 | 24
qa/158.out.2 | 2
qa/294.out | 32
qa/318.out | 2
qa/321.out | 2
qa/356.out | 4
qa/378.out | 14
qa/433.out | 2
qa/479.out.2 | 4
qa/501-linux.out | 114 -
qa/514.out.3 | 2
qa/519 | 16
qa/549.out | 2
qa/556.out | 2
qa/557.out | 4
qa/558.out.irix | 11
qa/558.out.linux | 11
qa/581.out | 16
qa/583 | 26
qa/583.out | 6
qa/622.out | 2
qa/668 | 412 ++++-
qa/668.out | 22
qa/733.out.2 | 2
qa/761.out | 18
qa/764.out | 4
qa/777.out | 2
qa/787.out | 2
qa/812 | 24
qa/812.out | 893 +++++------
qa/869 | 26
qa/869.out | 2902 ++++++++++++++++++++++++++++++++++++++
qa/922.out | 10
qa/951.out | 4
qa/983 | 112 +
qa/983.out | 49
qa/admin/check-vm | 2
qa/archives/simple-nvme.0 |binary
qa/archives/simple-nvme.index |binary
qa/archives/simple-nvme.meta |binary
qa/common.filter | 1
qa/group | 4
src/libpcp_pmcd/src/GNUmakefile | 7
src/pmdas/linux/proc_partitions.c | 16
src/pmdas/root/help | 1
src/pmdas/root/root.c | 5
src/pmdas/root/root.h | 1
src/pmdas/root/root_root | 1
src/pmie/src/dstruct.c | 16
src/pmie/src/dstruct.h | 13
src/pmie/src/eval.c | 11
src/pmie/src/eval.h | 3
src/pmie/src/fetch.sk | 7
src/pmie/src/pmie.c | 41
src/pmie/src/pragmatics.c | 164 --
src/pmie/src/pragmatics.h | 5
src/pmie/src/show.c | 83 -
src/pmie/src/symbol.c | 5
src/pmie/src/symbol.h | 5
src/pmie/src/syntax.c | 35
src/pmmgr/config/GNUmakefile | 12
src/pmmgr/pmmgr.cxx | 219 ++
src/pmmgr/pmmgr.h | 2
64 files changed, 4590 insertions(+), 876 deletions(-)
commit 7c679d92cd3da8e166c58b0b807bc526035e1f9b
Author: Nathan Scott <nathans@xxxxxxxxxx>
Date: Fri Oct 9 14:16:10 2015 +1100
pmmgr: remove unhelpful comments about libpcp and containers
If ever we support nested containers, its going to be pmdaroot
that needs to handle that, not libpcp. And in general, if you
are adding comments to code: please don't express a personal /
sarcastic opinion, and don't say "we should document X" (then
proceed to not do so). When someone does do these things, the
comments will be out of date and need fixing too, but are quite
likely to be forgotten since they're in unrelated code.
commit c67c67ef3f48211bc7d79d0d5def284fc5db00fd
Author: Nathan Scott <nathans@xxxxxxxxxx>
Date: Fri Oct 9 13:53:02 2015 +1100
build: FreeBSD build fix for 10.2 with dtrace probes auto-enabled
It seems FreeBSD needs to use the Solaris mode of running dtrace
helper tool - otherwise build failure results.
commit d3a01547e60855882efe8f11483fdcc089f370cf
Author: Nathan Scott <nathans@xxxxxxxxxx>
Date: Fri Oct 9 13:50:57 2015 +1100
pmdaroot: add a metric to aid mapping containers to their cgroups
Client tools need a way to map from the instances of the cgroup.*
metrics back to the containers for which those cgroups were built.
commit faf4c1b47912849e5a5b8233da58b9c2be760143
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Fri Oct 9 13:23:15 2015 +1100
qa/869: (new) check NVME disk names for archive from linux pmda
commit edd49eb87fe107f11c2fe3a91e50ea75628f4848
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Fri Oct 9 13:20:20 2015 +1100
linux pmda: add Intel copyright to proc_partions.c
As per email from Doug Nelson (doug.nelson@xxxxxxxxx) ...
The lawyer said we could use the same copyright line that everybody
else was using, so here's a new patch.
commit 452ec707498fb828bc4c049cc578e52e70856808
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Fri Oct 9 10:17:22 2015 +1100
qa/admin/check-vm: report correct Debian package for JSON Perl module
commit fc8f9b86e39bf29ebc073c427d2a8495ec8e1ebf
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Thu Oct 8 20:05:07 2015 +1100
pmie man page: minor wordsmithing around (new) %c format item
commit 3f9ebb0358790ac256ed18aeb8932979eed3b274
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Thu Oct 8 20:03:51 2015 +1100
src/pmie/show.c: small changes afer FQDN changes
In showSyn() should report m->hname not m->hconn for CND_FETCH
or CND_DELAY nodes in the expression tree to match the user's
input rule (remove question in comment).
Change type of findFormat() from int to char and add the default
case for %? (action is to keep searching, which was the implied
but is now explicit).
In showSatisfyingValue() revert
length = concat("\n<tab> ", length, &string);
to
length = concat("\n<space> ", length, &string);
commit d654b30fec41a1861d689c3c23ba38ff21b1f70c
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Thu Oct 8 20:02:18 2015 +1100
qa/294: track recent pmie changes
The number of context connections and disconnections is different
because there are now fewer pmNewContext() calls on the default
context path.
commit 062fda7abed17e2699ac0f8b763426ab5acccb4e
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Thu Oct 8 20:01:07 2015 +1100
qa/583: tweak filtering after recent pmie change
The hostname in the output lines are sensitive to DNS configuration on
the local and remote hosts where pmcd is running ... they may or may
not contain a FQDN ... adjust filter to accommodate both possibilities.
commit d19a4625acb8ae9f5b84529af2b7621d8535bca6
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Thu Oct 8 10:53:16 2015 +1100
qa/622: updated output after pmie report format change
commit 8214db66045487fdb39eedb7e4b925e6f1be74ea
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Thu Oct 8 10:49:33 2015 +1100
qa/519: tweak filtering after recent pmie change
The hostname in the output lines are sensitive to DNS configuration
on the remote hosts where pmcd is running ... they may or may not
contain a FQDN ... adjust filter to accommodate both possibilities.
commit ec5e592e7c514a542602fcf53042613fcb046536
Author: Frank Ch. Eigler <fche@xxxxxxxxxx>
Date: Sun Aug 23 11:55:33 2015 -0400
pmie-fqdn: support "%c" substitution in action strings
Add this peer to "%h" to expose the recently reified
connection-string value.
commit 48ddab879fc5dd3c2350de16dc3288dadc9fac4a
Author: Frank Ch. Eigler <fche@xxxxxxxxxx>
Date: Sun Aug 23 11:34:14 2015 -0400
pmie-fqdn qa/983: pmie -h test
New test for pmie fqdn functionality, to assert that it will only
connect to a remote host via the -h $CONNSTRING rather than "host
name". Done by suppressing normal pmie; starting at some arbitrary
high address, and for bonus marks also starting a socat relay proxy.
A brief pmie job against these (and a control scenario) is run.
commit 547cb50dd0b739988c81858e9a5346e8caed31c2
Author: Frank Ch. Eigler <fche@xxxxxxxxxx>
Date: Fri Mar 7 21:35:36 2014 -0500
PR1105: fqdn -h support for pmie
pmie's use of host names as interchangeable with pmNewContext
connection strings was deeply embedded. The new fields in structs
'metric' and 'host' track a connection string alongside the host name.
The newContext() function is now solely responsible for looking up
host names from archive or real-time pmcd connections, and updating
the pmie mmv data with it. Some other functions that used to take an
ambiguous name parameter now document which, or else take both.
pmie diagnostic messages try to more consistently print one or the
other, requiring many trivial qa/FOO.out updates.
In the man page, pmie's -H (don't do dns lookup) became undocumented,
since it has long been unimplemented. A little bit of clarification
regarding host names vs connection strings was added.
Some const qualifiers to char*'s have been added,
out of a sense of unchanging constervativeness.
commit 30e4fe45b73715a6f7ede02fbaf6ff592fe97410
Author: Suruchi Kadu <suruchi.a.kadu@xxxxxxxxx>
Date: Thu Oct 8 07:03:11 2015 +1100
linux pmda: add support for NVME devices
Patch from Intel.
Handle NVME device partition names like nvme0n1p1 using the
same recipe as the other SSD ones, "ram...", "mmcblk...", etc.
commit 5328a95b3f9dae7aa952f791736b044a565ea013
Author: Lukas Berk <lberk@xxxxxxxxxx>
Date: Wed Oct 7 13:16:13 2015 -0400
Modify qa/812 to use hostzone options (fixes buildbot timezone based
failures)
Buildbot tests were failing 812 due to having several hour differences
in the dump/extract output. Appending '-z' output fixes this, along
with updating the output for the testcase.
commit 1f09b70af7d856eebd6d54451fbb211906a97fe2
Author: Frank Ch. Eigler <fche@xxxxxxxxxx>
Date: Tue Oct 6 18:43:12 2015 -0400
pmmgr container pmie qa: track expected number of pmie.log files
commit 93a76c70157a501f7d21c08766931b42b05484c1
Author: Nathan Scott <nathans@xxxxxxxxxx>
Date: Tue Oct 6 17:09:50 2015 -0400
pmmgr PR1105: enable pmie for subtarget-containers
commit be079febc8bfc6744e440f40be074272636c9eaf
Author: Frank Ch. Eigler <fche@xxxxxxxxxx>
Date: Sun Apr 5 17:17:34 2015 -0400
pmmgr subtarget-containers: disable by default
Investigation of PR1106 indicates that pmcd is too fragile to use with
containers safely, e.g., a pmda can get stuck within a container
namespace, rendering further normal use of that pmda useless. Until
this is fixed, let's not supply /etc/pcp/pmmgr/subtarget-containers.
commit 5410ba6d2b4fd2f1d09048e3f3f4704c85d450b2
Author: Frank Ch. Eigler <fche@xxxxxxxxxx>
Date: Wed Apr 1 21:05:21 2015 -0400
pmmgr: lengthen subtarget-containers qa/668
Further testing on a slower virtual machine shows situations where
pmlogconf & docker run dramatically slower than on full workstations,
whether due to bare-metalness, number of processors, presence of
selinux bugs & troubleshooting thereof, etc. This test case is
made more conservative in terms of timing (3ish minutes) and more
defensive of iffy docker situations.
commit f1b8bb9794b07c967925038c6c68e9463b2a931f
Author: Frank Ch. Eigler <fche@xxxxxxxxxx>
Date: Tue Mar 31 21:15:23 2015 -0400
pmmgr: subtarget-containers qa/668
A relatively short (60-second) test that fires up a pmmgr only in
order to attach pmloggers to a few *real live* docker containers.
(That means the qa-invoking user probably needs to be in group
'docker'; something checked by the qa script.) The docker container
being used is a tiny "busybox sleep XX" kind, so there should be
only minimal disk/network consumption.
The tests do not assert anything about the contents of the pcp
archives resulting from the per-container attachment. In my tests,
they got apprx. zero actual content, other than proper pmcd.hostname.
There ought to be a pmlogconf ditty for the container's own cgroup
stats, and disk/network stuff should eventually be correct.
commit 4e0017d4ab85626fe22d0107cd0a8b75cbcf5953
Author: Frank Ch. Eigler <fche@xxxxxxxxxx>
Date: Tue Mar 31 18:35:23 2015 -0400
pmmgr: subtarget-containers support
Add support for a config option "subtarget-containers", which
implicitly forks further pmlogger (but not pmie: PR1105) instances for
all running containers on an identified target host. The host's
host-id string gets a "--CONTAINER" code attached, and the connection
hostspec gets a "container=XXXXX" appended with a ? or & suffix-prefix.
commit 583b350d98246f9082253f04b76b682985c0de55
Author: Frank Ch. Eigler <fche@xxxxxxxxxx>
Date: Sun Apr 5 17:17:34 2015 -0400
pmmgr subtarget-containers: disable by default
Investigation of PR1106 indicates that pmcd is too fragile to use with
containers safely, e.g., a pmda can get stuck within a container
namespace, rendering further normal use of that pmda useless. Until
this is fixed, let's not supply /etc/pcp/pmmgr/subtarget-containers.
commit 35c490f3f6ca7c800a09d1903bfc2a9e03a83352
Author: Frank Ch. Eigler <fche@xxxxxxxxxx>
Date: Wed Apr 1 21:05:21 2015 -0400
pmmgr: lengthen subtarget-containers qa/668
Further testing on a slower virtual machine shows situations where
pmlogconf & docker run dramatically slower than on full workstations,
whether due to bare-metalness, number of processors, presence of
selinux bugs & troubleshooting thereof, etc. This test case is
made more conservative in terms of timing (3ish minutes) and more
defensive of iffy docker situations.
commit d42afe606b02e2fa570b34a0313add8aa555619a
Author: Frank Ch. Eigler <fche@xxxxxxxxxx>
Date: Tue Mar 31 21:15:23 2015 -0400
pmmgr: subtarget-containers qa/668
A relatively short (60-second) test that fires up a pmmgr only in
order to attach pmloggers to a few *real live* docker containers.
(That means the qa-invoking user probably needs to be in group
'docker'; something checked by the qa script.) The docker container
being used is a tiny "busybox sleep XX" kind, so there should be
only minimal disk/network consumption.
The tests do not assert anything about the contents of the pcp
archives resulting from the per-container attachment. In my tests,
they got apprx. zero actual content, other than proper pmcd.hostname.
There ought to be a pmlogconf ditty for the container's own cgroup
stats, and disk/network stuff should eventually be correct.
commit 8864f4e2d9824933d8d598e8361869d58815a2b8
Author: Frank Ch. Eigler <fche@xxxxxxxxxx>
Date: Tue Mar 31 18:35:23 2015 -0400
pmmgr: subtarget-containers support
Add support for a config option "subtarget-containers", which
implicitly forks further pmlogger (but not pmie: PR1105) instances for
all running containers on an identified target host. The host's
host-id string gets a "--CONTAINER" code attached, and the connection
hostspec gets a "container=XXXXX" appended with a ? or & suffix-prefix.
|