[pcp] Dynamic PMNS changes finally arrive
Mark Goodwin
goodwinos at gmail.com
Mon Sep 28 19:59:11 CDT 2009
Ken McDonell wrote:
> Sounds like a plan.
>
> If you're planning on running QA after merging my changes, you'll
> probably need my QA ... should I just go ahead and push those changes
> into the pcpqa tree on oss.sgi.com?
might as well, perhaps commit a tag first, see git-tag(1),
something like "prior to dynamic pmns".
What QA platforms h/w and s/w are you running on?
Cheers
- Mark
>
> On Tue, 2009-09-29 at 09:13 +1000, Mark Goodwin wrote:
>> Hi Ken,
>>
>> I have a round of changes (v3.0.0-6) to push up to the dev branch
>> (clean up "the mess" following the Fedora keel-haul, and also
>> integrates the perl packaging) - this is under QA at the moment,
>> which is taking longer than I'd hoped.
>>
>> In any case, I'll push 3.0.0-6 to 'dev' asap. Then I'll merge your
>> master branch and run a QA delta. If that goes OK I propose we push
>> this to (v3.0.0-7) for review and soak time in the dev branch since
>> the design is already reviewed. At the very least we'll need to
>> update the cluster PMDA before 3.0 can be considered release worthy.
>>
>> Cheers
>> -- Mark
>>
>> Ken McDonell wrote:
>>> Committed to the *master* branch of my PCP tree on oss.sgi.com
>>>
>>> There are several rounds of commits for this set of changes that became
>>> much more complicated and touched more files than I'd originally
>>> expected.
>>>
>>> This all passes my QA ... after I get some feedback on the PCP changes
>>> I'll move the QA tree forward.
>>>
>>> In the logs below there are lots of changes from the *master* and *dev*
>>> branches (this is a mess we need to sort out pronto!). Only the ones
>>> authored by me are really part of this batch of commmits, the others are
>>> innocent bystanders as a result of the passage of time.
>>>
>>> Changes committed to git://oss.sgi.com/kenj/pcp.git master
>>>
>>> .gitignore | 2
>>> CHANGELOG | 33 +
>>> VERSION.pcp | 10
>>> build/GNUmakefile | 3
>>> build/rpm/pcp.spec.in | 39 +-
>>> build/rpm/pcp_fedora.spec | 385 +++++++++++++++++-----
>>> build/sun/GNUmakefile | 53 ++-
>>> configure.in | 99 ++++-
>>> debian/changelog | 21 +
>>> debian/control | 72 ++--
>>> man/man1/GNUmakefile | 1
>>> man/man1/newhelp.1 | 24 +
>>> man/man1/pmcd.1 | 6
>>> man/man1/pmdaib.1 | 113 ++++++
>>> man/man1/pmval.1 | 27 +
>>> man/man3/pmda.3 | 15
>>> man/man3/pmdachildren.3 | 139 +++++++-
>>> man/man3/pmdamain.3 | 84 +++-
>>> man/man3/pmdaname.3 | 91 +++++
>>> man/man3/pmdapmid.3 | 89 ++++-
>>> man/man4/pmns.4 | 45 +-
>>> src/cpan/MMV/Changes | 7
>>> src/cpan/MMV/MMV.pm | 6
>>> src/cpan/MMV/MMV.xs | 41 +-
>>> src/cpan/MMV/server.pl | 17
>>> src/cpan/PMDA/Changes | 8
>>> src/cpan/PMDA/PMDA.pm | 6
>>> src/cpan/PMDA/PMDA.xs | 39 +-
>>> src/cpan/PMDA/local.c | 3
>>> src/dbpmda/src/GNUmakefile | 2
>>> src/dbpmda/src/dbpmda.c | 64 +++
>>> src/dbpmda/src/dbpmda.h | 2
>>> src/dbpmda/src/dso.c | 205 +++++++++--
>>> src/dbpmda/src/gram.y | 43 ++
>>> src/dbpmda/src/lex.l | 30 -
>>> src/dbpmda/src/pmda.c | 130 ++++++-
>>> src/dbpmda/src/util.c | 20 +
>>> src/genpmda/genpmda | 4
>>> src/include/builddefs.in | 13
>>> src/include/impl.h | 31 +
>>> src/include/mmv_stats.h | 2
>>> src/include/pmda.h | 73 +++-
>>> src/libpcp/src/.gitignore | 3
>>> src/libpcp/src/desc.c | 11
>>> src/libpcp/src/fetchlocal.c | 28 +
>>> src/libpcp/src/help.c | 11
>>> src/libpcp/src/instance.c | 46 +-
>>> src/libpcp/src/p_pmns.c | 2
>>> src/libpcp/src/pmns.c | 139 ++++++--
>>> src/libpcp/src/store.c | 14
>>> src/libpcp/src/util.c | 34 +
>>> src/libpcp_gui/src/.gitignore | 3
>>> src/libpcp_gui/src/GNUmakefile | 2
>>> src/libpcp_mmv/src/.gitignore | 1
>>> src/libpcp_mmv/src/mmv_stats.c | 65 ++-
>>> src/libpcp_pmcd/src/.gitignore | 1
>>> src/libpcp_pmda/src/.gitignore | 1
>>> src/libpcp_pmda/src/callback.c | 45 ++
>>> src/libpcp_pmda/src/help.c | 2
>>> src/libpcp_pmda/src/libdefs.h | 3
>>> src/libpcp_pmda/src/mainloop.c | 261 ++++++++++++---
>>> src/libpcp_pmda/src/open.c | 133 ++++---
>>> src/libpcp_trace/src/.gitignore | 3
>>> src/newhelp/chkhelp.c | 6
>>> src/newhelp/newhelp.c | 50 ++
>>> src/pmcd/GNUmakefile | 5
>>> src/pmcd/migrate_pcp_var_dir | 126 -------
>>> src/pmcd/pmdaproc.sh | 242 +++++++------
>>> src/pmcd/rc_pcp | 14
>>> src/pmcd/src/config.c | 163 +++++++--
>>> src/pmcd/src/dofetch.c | 57 ++-
>>> src/pmcd/src/dopdus.c | 660 ++++++++++++++++++++++++++++++++------
>>> src/pmcd/src/dostore.c | 14
>>> src/pmcd/src/pmcd.h | 8
>>> src/pmcd/src/pmcd.log | 198 +++++------
>>> src/pmcd/src/sample.log | 4
>>> src/pmcd/src/sendmail.log | 4
>>> src/pmcd/src/simple.log | 4
>>> src/pmcd/src/trace.log | 4
>>> src/pmdas/GNUmakefile | 7
>>> src/pmdas/ib/daemon.c | 20 -
>>> src/pmdas/ib/dso.c | 2
>>> src/pmdas/ib/ib.c | 283 +++++++++-------
>>> src/pmdas/ib/ibpmda.h | 4
>>> src/pmdas/ib/pmda.c | 4
>>> src/pmdas/linux/proc_cpuinfo.c | 38 --
>>> src/pmdas/linux/proc_cpuinfo.h | 1
>>> src/pmdas/linux/proc_loadavg.c | 7
>>> src/pmdas/mmv/.gitignore | 5
>>> src/pmdas/mmv/GNUmakefile | 28 -
>>> src/pmdas/mmv/Install | 6
>>> src/pmdas/mmv/Remove | 5
>>> src/pmdas/mmv/mmv.c | 103 +++--
>>> src/pmdas/mysql/pmdamysql.pl | 3
>>> src/pmdas/named/.gitignore | 2
>>> src/pmdas/named/GNUmakefile | 45 ++
>>> src/pmdas/named/Install | 32 +
>>> src/pmdas/named/Remove | 25 +
>>> src/pmdas/named/pmdanamed.pl | 155 ++++++++
>>> src/pmdas/pdns/.gitignore | 2
>>> src/pmdas/pdns/GNUmakefile | 45 ++
>>> src/pmdas/pdns/Install | 27 +
>>> src/pmdas/pdns/Remove | 25 +
>>> src/pmdas/pdns/pmdapdns.pl | 194 ++++++++++-
>>> src/pmdas/pmcd/src/pmcd.c | 12
>>> src/pmdas/samba/.gitignore | 2
>>> src/pmdas/samba/GNUmakefile | 49 ++
>>> src/pmdas/samba/Install | 43 ++
>>> src/pmdas/samba/Remove | 25 +
>>> src/pmdas/samba/pmdasamba.pl | 205 +++++++++++
>>> src/pmdas/sample/help | 24 +
>>> src/pmdas/sample/pmns | 1
>>> src/pmdas/sample/src/pmda.c | 2
>>> src/pmdas/sample/src/sample.c | 362 ++++++++++++++++++--
>>> src/pmdas/simple/pmdasimple.pl | 6
>>> src/pmdas/solaris/GNUmakefile | 15
>>> src/pmdas/solaris/common.h | 14
>>> src/pmdas/solaris/data.c | 212 +++++++++++-
>>> src/pmdas/solaris/help | 95 +++++
>>> src/pmdas/solaris/netmib2.c | 344 +++++++++++++++++++
>>> src/pmdas/solaris/netmib2.h | 62 +++
>>> src/pmdas/solaris/pmns.network | 31 +
>>> src/pmdas/solaris/pmns.zfs | 15
>>> src/pmdas/solaris/pmns.zpool | 20 +
>>> src/pmdas/solaris/root | 5
>>> src/pmdas/solaris/solaris.c | 7
>>> src/pmdas/solaris/zfs.c | 164 ++++++++-
>>> src/pmdas/solaris/zpool.c | 226 ++++++++++---
>>> src/pmdas/trace/GNUmakefile | 3
>>> src/pmdas/trace/GNUmakefile.demos | 2
>>> src/pmdas/trace/GNUmakefile.stub | 2
>>> src/pmdas/windows/fetch.c | 91 ++---
>>> src/pmdas/windows/libpdh.def | 1
>>> src/pmdas/windows/libpdh.h | 20 +
>>> src/pmdas/windows/open.c | 13
>>> src/pmdas/windows/pmda.c | 32 +
>>> src/pmdas/windows/pmns.kernel | 1
>>> src/pmie/rc_pmie | 12
>>> src/pmie/src/GNUmakefile | 2
>>> src/pmlc/GNUmakefile | 2
>>> src/pmlogextract/GNUmakefile | 2
>>> src/pmlogger/GNUmakefile | 2
>>> src/pmlogger/pmlogger.c | 3
>>> src/pmns/Make.stdpmid | 8
>>> src/pmns/ReplacePmnsSubtree | 8
>>> src/pmns/pmnsutil.c | 5
>>> src/pmns/stdpmid.pcp | 8
>>> src/pmpost/GNUmakefile | 2
>>> src/pmproxy/rc_pmproxy | 12
>>> src/pmval/pmval.c | 55 ++-
>>> 150 files changed, 6178 insertions(+), 1581 deletions(-)
>>>
>>> commit 52872f6b735fbf7814443f226df21eee30b35936
>>> Author: Ken McDonell <kenj at internode.on.net>
>>> Date: Sun Sep 27 16:30:09 2009 +1000
>>>
>>> ReplacePmnsSubtree fix bad signal handling botch
>>>
>>> The trap handling was totally wrong in cleanup() ... causing the
>>> command "1" to be executed instead of the null command "".
>>>
>>> Exposed by another bug in a related area ... if the PMNS lock
>>> exists already, after the timeout in pmlock we fall into the
>>> trap cleanup() routine instead of charing on, so the script
>>> fails horribly (and leaves the stale lock file around).
>>>
>>> Also fixed an unlikely buffer over-run in mmv.c.
>>>
>>> commit 4087051b86b580d7c7cb6e50806edb4eaa63e4d4
>>> Author: Ken McDonell <kenj at internode.on.net>
>>> Date: Sun Sep 27 15:20:24 2009 +1000
>>>
>>> newhelp, dbpmda, libpcp, libpcp_pmda, pmcd, sample pmda - dynamic PMNS changes
>>>
>>> This is the final round of changes that get all of the changes for
>>> PMDA_INTERFACE_4 and the support of dynamic subtress of the PMNS where
>>> the PMDA (not PMCD) maintains knowledge of the PMNS.
>>>
>>> With these changes and a related round of QA changes, we're passing all
>>> of the QA tests (excluding the remote group that I cannot test at the
>>> moment, but are unlikely to find problems in this code).
>>>
>>> commit 40e120d100d8f8f2a638124edef1ac2a16c0b975
>>> Author: Ken McDonell <kenj at internode.on.net>
>>> Date: Sun Sep 27 15:18:33 2009 +1000
>>>
>>> pmdaproc.sh Revert changes for perl pmda setup.
>>>
>>> Just put it back the way it was before I started messing with it!
>>>
>>> commit 4cead6907bf315ed009725db26659d3c6e12c698
>>> Author: Ken McDonell <kenj at internode.on.net>
>>> Date: Sun Sep 27 15:11:44 2009 +1000
>>>
>>> pmdaproc.sh tweaking to improve QA success
>>>
>>> Add delays after sending pmcd SIGHUP and before checking for pmcd health
>>> to avoid small window where badness happens to pmcd from new PMDA, but
>>> after the test has been done.
>>>
>>> Also wound back previous Perl PMDA setup change as it does not help the
>>> QA test for the news PMDA ... reworked the QA test instead!
>>>
>>> commit 9371a85a8a5b65b43be732e585e1b8f98f4b94e1
>>> Author: Ken McDonell <kenj at internode.on.net>
>>> Date: Thu Sep 24 16:59:07 2009 +1000
>>>
>>> Only remake pmns file if it does not exist.
>>>
>>> Part of the perl PMDA support was unconditionally remaking pmns and
>>> domain.h ... this was reordering the metrics for the simple PMDA and
>>> causing QA tests 255 and 578 to fail ... more importantly if the
>>> pmns file already exists then there is no point in using the perl
>>> PMDA magic to recreate it.
>>>
>>> commit 4f6a801b054f92b23064385940a441bc66b758bd
>>> Merge: e421a49... 11549ac...
>>> Author: Ken McDonell <kenj at internode.on.net>
>>> Date: Thu Sep 17 20:00:33 2009 +1000
>>>
>>> Merge branch 'master' of git://oss.sgi.com/kenj/pcp
>>>
>>> Conflicts:
>>> CHANGELOG
>>> VERSION.pcp
>>> src/pmcd/rc_pcp
>>> src/pmie/rc_pmie
>>> src/pmproxy/rc_pmproxy
>>>
>>> commit 11549aca86588c4695d26829e02f2ec9e7b43c3d
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Mon Sep 7 16:29:34 2009 +1000
>>>
>>> Bump pcp stable version to 2.9.2.
>>>
>>> commit 7997eea93ad9c240d2cbc7c2bfbd9ae21bb29e35
>>> Author: Ken McDonell <kenj at internode.on.net>
>>> Date: Mon Sep 7 11:59:21 2009 +1000
>>>
>>> Fix the number-of-samples value calculated in pmval once and for all.
>>>
>>> For those who care, it turns out that the number of _fetches_ may be
>>> known, the number of _reported_ samples cannot be precomputed correctly
>>> in all cases ...
>>> - only works for archives with -a (not -U or -h)
>>> - is known for non-counter metrics
>>> - most of the time is known for counter metrics but there are some
>>> important corner cases where this is not the case, specifically there
>>> will be one more reported sample than expected for a counter when the
>>> report starting time is not at the start of the archive and the metric
>>> has a defined value before the report starting time with no intervening
>>> mark record (interpolate mode does return a value at the first report
>>> sample in this case!).
>>>
>>> With this pmval patch, the QA fallout is nil, except for 144 that is an
>>> unrelated timezone problem. Also this patch avoids referencing the
>>> context too soon, as there is no point in inspecting the metric's
>>> semantics to try and better guess the value of smpls. As a bonus, this
>>> patch also documents pmval's -U option that has remained a secret for
>>> nearly a decade.
>>>
>>> commit f4e6d55fd0466a706144ca61b4428436ecea7eb7
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Mon Sep 7 11:51:58 2009 +1000
>>>
>>> Attempt to resolve sys/stat related build issue on Alpha.
>>>
>>> Not having access to an Alpha makes this difficult, but one thing
>>> that is clearly not helping is that with Linux glibc on Alpha, we
>>> end up defining *both* HAVE_STAT_TIMESPEC and HAVE_STAT_TIME_T, &
>>> that breaks otherwise valid pmns.c coding assumptions.
>>>
>>> This change means only one type for the stat time fields can be
>>> defined, and it errs on the side of timespec (which should be the
>>> one Alpha uses, AFAICT) - so perhaps this issue will finally be
>>> resolved and PCP can progress in the Debian build/release system.
>>>
>>> commit be8793c58935362184a1dd4351388ca6a0013f3b
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Mon Sep 7 11:26:09 2009 +1000
>>>
>>> Fix mmv library build on mac/win - missing a header.
>>>
>>> commit 92eb7b896ff6eb83b2feef50e6a89c74c3aa1717
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Fri Sep 4 14:17:37 2009 +1000
>>>
>>> Bump pcp stable version to 2.9.1.
>>>
>>> commit 2e890f3d1d9c4af31fc22af95cb7e0f20336523e
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Fri Sep 4 14:15:16 2009 +1000
>>>
>>> Correct pmval sample count calculation.
>>>
>>> Martin reported several QA tests were failing wrt pmval "samples" line.
>>>
>>> What I've found is the problem either exists/doesn't exist depending on
>>> several factors:
>>> - running on a 32 bit machine, Debian unstable, the problem never occurs
>>> no matter how pmval is built.
>>> - running on a 64 bit machine, Debian unstable, the problem only happens
>>> with the packaged binaries. The key factor appears to be the introduction
>>> of -O2, which looks like it reliably changes certain values on the stack,
>>> that would otherwise be zero (used uninitialised, in either case).
>>>
>>> There's two problems here, I believe:
>>> - in pmval.c the getargs() call initialises the value of "smpls" (number
>>> of samples to take) based on cntxt->desc.sem. However, at that point,
>>> we have not established a context yet, and so that field is uninitialised.
>>> - this test doesn't match the comment above it, and doesn't make sense:
>>>
>>> /* counters require 2 samples to produce reported sample */
>>> if (*smpls > 0 && cntxt->desc.sem != PM_SEM_COUNTER)
>>>
>>> that would seem to me that it should be "cntxt->desc.sem == PM_SEM_COUNTER".
>>>
>>> commit 6940bbf89ca10d395fff93eeb7ce7424cec5d2ad
>>> Merge: b7bb608... 350246f...
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Fri Sep 4 14:08:24 2009 +1000
>>>
>>> Merge branch 'dev' of git://oss.sgi.com/nathans/pcp into dev
>>>
>>> commit 350246f0ba359876660994823d2e37b864040c3f
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Thu Sep 3 12:59:57 2009 +1000
>>>
>>> Another pass at getting the perl/debian packaging auto-deps right.
>>>
>>> commit 91c9f4a0a06d8e300c23f49feab09141c488b0f5
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Thu Sep 3 12:59:24 2009 +1000
>>>
>>> Correct mmv install targets makefile macro.
>>>
>>> commit bf33333edbcf48ee84dad07cace240d6aff6cc99
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Wed Sep 2 15:07:19 2009 +1000
>>>
>>> Workaround another oddity in (very) recent versions of autoconf.
>>>
>>> Problem found by the Debian autobuilders, in the pcp-gui build but it
>>> affects pcp too. Latest versions of autoconf evaluate AC_PROG_LEX in
>>> the compiler check macro, and complete remove any trace of AC_PROG_LEX
>>> from its original callsite. This results in a syntax error inside our
>>> configure script, as we call it within if ... fi, which then evaluates
>>> to having no body to the if. Ugh. Fix is trivial though, since these
>>> macros do the user-env-variable expansion thing already.
>>>
>>> commit efcb3b590f086feb1507f82716d2f0b6eee69d33
>>> Author: Petter Reinholdtsen <pere at hungry.com>
>>> Date: Wed Sep 2 14:38:34 2009 +1000
>>>
>>> Add missing dependencies and runlevels into PCP start scripts.
>>>
>>> With dependency based boot sequencing, I discovered what I believe is
>>> a bug in the init.d scripts of this package. The list of runlevels to
>>> stop in is incomplete. The script should stop in runlevel 1 to make
>>> sure the service is started again when switching from runlevel 1 to
>>> 2-5. Also, the script should probably stop during halt (0) and reboot
>>> (6). The package should start in runlevel 4, just like in runlevels
>>> 2, 3 and 5. I suspect the list of dependencies are incomplete. The
>>> scripts used to start after the syslog collector was operational, and
>>> should probably continue to do so. Also, the scripts need files in
>>> /usr/ and should depend on $remote_fs.
>>>
>>> If the scripts themselves need to start in any specific order, the
>>> relationship should be listed explicitly in the LSB header too. Make
>>> sure the update-rc.d calls in the postinst reflect this ordering if
>>> any such dependencies should exist.
>>>
>>> This patch implement the proposed change. Without it, the script seem
>>> to fail to handle runlevel 0, 1, 4 and 6 properly, and might start too
>>> early in the boot.
>>>
>>> commit 4aa31ed00a60c3926a53bbd8d47aa0b2eaf254c4
>>> Merge: 35c230e... 3323bb9...
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Wed Sep 2 13:18:44 2009 +1000
>>>
>>> Merge branch 'dev' of git://oss.sgi.com/nathans/pcp into dev
>>>
>>> commit b7bb60801c2fc342b15473f1e3b341dd8ec7b84a
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Wed Sep 2 12:38:19 2009 +1000
>>>
>>> Update gitignore files for Mac and Windows library file names.
>>>
>>> commit 3323bb90ad5f66480c05b7fe182a72a0813c9b41
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Wed Sep 2 12:30:47 2009 +1000
>>>
>>> Fix a segv observed in the Windows PMDA accessing SQL Server metrics.
>>>
>>> The recent optimisation added for direct instance domain lookup added
>>> a flaw that could case an array index of -1 to be used to access into
>>> the metric value table. Observed on SQL Server 2008 & Windows Server
>>> 2008 R2 only so far.
>>>
>>> commit 1112ac636cae6feac624279c76239003662d4a35
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Wed Sep 2 11:49:20 2009 +1000
>>>
>>> Small MMV pmda changes to allow DSO builds in the future.
>>>
>>> Turns out that the current protocol for getting pmcd to
>>> update its namespace cannot be used from DSO agents ...
>>> one more reason to revamp this area.
>>>
>>> commit 677c6f2cea7fce4ff5f10f939fc5a942861719dd
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Wed Sep 2 11:46:40 2009 +1000
>>>
>>> Fix logic error causing wrong windows version to be reported.
>>>
>>> commit e421a49323307d49f6b897ae87f94c934b6fdcad
>>> Author: Ken McDonell <kenj at internode.on.net>
>>> Date: Tue Sep 1 17:44:22 2009 +1000
>>>
>>> Interim commit for start of daemon PMDA support for dynamic PMNS.
>>>
>>> commit 210be7f679b10d71dda985fdebc7dd9aec5d606a
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Fri Aug 28 08:44:50 2009 +1000
>>>
>>> Correct some regular expressions in Samba PMDA.
>>>
>>> Of particular note, some of the heading lines in the
>>> smbstatus --profile output have trailing whitespace.
>>> Otherwise, this PMDA is now fully functional; and as
>>> it used the stock Samba utilities, is not dependent
>>> on the many flavours of Samba shmem format anymore.
>>> At the cost of a slight overhead increase, of course,
>>> but the original was unmaintainable.
>>>
>>> Thanks to James Peach for advise on redesigning this.
>>>
>>> commit 0758690bcbbf26db438560d5b5b94b1644979047
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Thu Aug 27 14:26:32 2009 +1000
>>>
>>> Descend into samba PMDA during the build (SUBDIRS update).
>>>
>>> commit 9e758dd286ee2c49ecd991599a4c87a5520b9892
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Thu Aug 27 14:25:18 2009 +1000
>>>
>>> Add kernel.all.uptime to the Windows kernel agent.
>>>
>>> This code uses the PdhGetFormatterCounterValue API, which we
>>> have not used before on Windows, hence the changes to libpdh
>>> header and dll definition files.
>>>
>>> commit b0cbeab4df3aec93a2f9749c7912dc8adc0b27ea
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Thu Aug 27 10:21:28 2009 +1000
>>>
>>> Rearrange platform-specific handling in pmdaproc.sh.
>>>
>>> Keep everything that does platform-specific stuff in one
>>> place now, run early on, instead of all over the shop.
>>>
>>> The code that handles determining whether the PCP::PMDA
>>> Perl module is installed is also reworked to not fail in
>>> the case where perl_opt is set, but user doesn't select
>>> that option when Install'ing a PMDA. This bit Mark with
>>> the simple PMDA, which has both Perl and C variants.
>>>
>>> commit ac9170cf888639cc731474d39cb100e9d7b2c519
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Thu Aug 27 09:20:14 2009 +1000
>>>
>>> Initial version of the (perl-based) Samba PMDA.
>>>
>>> commit e00aafe76a538ad09322499c2f669450d3b0c486
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Wed Aug 26 11:46:12 2009 +1000
>>>
>>> Fix inconsistent metric name in pmdasimple.pl - simple.now was wrong.
>>>
>>> commit e937d8332c9dc6a0f4d5ea5fb10b6d3530b7261b
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Wed Aug 26 11:02:55 2009 +1000
>>>
>>> Remove hard-coded path in the perl variant of the simple PMDA (thanks Mark).
>>>
>>> commit 531de7440179d323c8774fad2a4c13ac4c21bcbb
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Wed Aug 26 11:00:00 2009 +1000
>>>
>>> Use PM_IN_NULL in pmdasimple.pl so it functions correctly on all platforms.
>>>
>>> commit 956be698fe2de60b96c2c6fa72be1555a1c3cf7c
>>> Merge: 7bf0f27... b8cb2dd...
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Mon Aug 24 10:58:57 2009 +1000
>>>
>>> Merge branch 'dev' of git://oss.sgi.com/nathans/pcp into dev
>>>
>>> commit 7bf0f27d27c11db4002699e74611ef718551add3
>>> Merge: 309e7be... 3db8e4f...
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Mon Aug 24 10:57:30 2009 +1000
>>>
>>> Merge branch 'solaris' of git://oss.sgi.com/makc/pcp into dev
>>>
>>> commit 309e7be90719ebc12bc66e9a06999909e7f66c16
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Mon Aug 24 09:03:19 2009 +1000
>>>
>>> Add PCP::PMDA helper routines for determining native long sizes.
>>>
>>> commit 90e8cd865a6bfabf91315feea3743054a1511111
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Mon Aug 24 09:01:38 2009 +1000
>>>
>>> Extend MMV to allow teardown of MMV file, and set errno on failure.
>>>
>>> Setting errno helps alot with diagnosing mmv_stats_init() failures.
>>>
>>> commit d4732eb5d60d8d31bb6eba69daa2787befb34325
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Mon Aug 24 08:56:24 2009 +1000
>>>
>>> Initial stable (1.00) version of MMV Perl interface.
>>>
>>> Changes include:
>>> - added stop interface
>>> - added stats_set interface
>>> - fixed incorrect memory free on init
>>> - extend server.pl example to report idle time
>>> - fixed use of av_len return codes (off by one)
>>>
>>> commit 3db8e4f3626241cb40e52f36a691263c8aaa34ff
>>> Author: Max Matveev <makc at gmx.co.uk>
>>> Date: Mon Aug 17 18:52:49 2009 +1000
>>>
>>> Release zpool resources when finished updates
>>>
>>> It's not documented but it's obvious that zpool_iter expects the
>>> callback to call zpool_close to release resources associated with
>>> the zpool - it's not a problem for short-lived commands like zpool(1)
>>> but it leaks memory like a firehose.
>>>
>>> commit 35c230ef34a191bbb5b8f12b91d122e2863ac740
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Fri Aug 14 14:58:10 2009 +1000
>>>
>>> Correct the metric units for several Windows per-process memory metrics.
>>>
>>> commit 758b2b7039e878189b42fab6d8d3a789a27810d1
>>> Author: Max Matveev <makc at gmx.co.uk>
>>> Date: Wed Aug 12 00:01:59 2009 +1000
>>>
>>> Eliminate incidents of crime against whitespace
>>>
>>> commit 012af5fa4e16029513c6d1c2ff7c7a3d4a3069e3
>>> Author: Max Matveev <makc at gmx.co.uk>
>>> Date: Tue Aug 11 23:55:59 2009 +1000
>>>
>>> Hide instances for destroyed pools and filesystems
>>>
>>> In Solaris PMDA if a zpool or a zfs is removed make sure that the
>>> corresponding instance disappears from the list of instances returned on
>>> fetch.
>>>
>>> commit a8cbf7b86b1bab119246006b23b1114fe1b4b9d7
>>> Author: Ken McDonell <kenj at internode.on.net>
>>> Date: Fri Aug 7 11:22:23 2009 +1000
>>>
>>> Revert __pmProcessTerminate() change.
>>>
>>> Since perl PMDAs are now all running in their own process group revert
>>> to allow kill(0, ...) to proceed.
>>>
>>> commit 8ea858d245d093ed3285e0e4370303bd678cc468
>>> Merge: 9b4b935... ead91ff...
>>> Author: Ken McDonell <kenj at internode.on.net>
>>> Date: Thu Aug 6 18:35:49 2009 +1000
>>>
>>> Merge branch 'dev' into pmns
>>>
>>> commit 9b4b9351a7e41bc21f893063e0311accd7cee046
>>> Author: Ken McDonell <kenj at internode.on.net>
>>> Date: Thu Aug 6 18:18:10 2009 +1000
>>>
>>> Guard in _pmProcessTerminate() to prevent killing all member of process group.
>>>
>>> Problem exposed by qa/162 and perl implementation of news pmda.
>>>
>>> commit 7f72e4acff1b876b10e6f6432a55de2e49b9e148
>>> Author: Ken McDonell <kenj at internode.on.net>
>>> Date: Thu Aug 6 18:16:49 2009 +1000
>>>
>>> Second round of changes for PMDA_INTERFACE_4 dynamic metrics.
>>>
>>> DSO mode PMDAs (sampledso) is very close to working correctly now.
>>>
>>> commit ead91ffca2605502c3552ffbc228ffbc0ed3ecc3
>>> Author: Mark Goodwin <mgoodwin at redhat.com>
>>> Date: Thu Aug 6 12:14:25 2009 +1000
>>>
>>> update changelog and delete unncessary Vendor string
>>>
>>> Signed-off-by: Mark Goodwin <mgoodwin at redhat.com>
>>>
>>> commit b77cb57bb669b1129c1241f7818fe75935390a3b
>>> Author: Mark Goodwin <mgoodwin at redhat.com>
>>> Date: Thu Aug 6 12:06:23 2009 +1000
>>>
>>> also bump the src tarball version
>>>
>>> Signed-off-by: Mark Goodwin <mgoodwin at redhat.com>
>>>
>>> commit 6c8311051895092662b8ee1986945a3846577b04
>>> Author: Mark Goodwin <mgoodwin at redhat.com>
>>> Date: Thu Aug 6 12:01:24 2009 +1000
>>>
>>> bump to 3.0.0-2
>>>
>>> Signed-off-by: Mark Goodwin <mgoodwin at redhat.com>
>>>
>>> commit 99aa1c21c22e64350ba750284039653750633540
>>> Merge: ef9852c... f2866c9...
>>> Author: Mark Goodwin <mgoodwin at redhat.com>
>>> Date: Thu Aug 6 11:37:29 2009 +1000
>>>
>>> merge pcp-2.9.0-1 dev
>>>
>>> modified: CHANGELOG
>>> modified: build/rpm/pcp.spec.in
>>> modified: debian/changelog
>>> modified: debian/control
>>> modified: src/cpan/PMDA/PMDA.xs
>>> modified: src/libpcp_pmda/src/open.c
>>> modified: src/pmcd/src/config.c
>>> modified: src/pmdas/GNUmakefile
>>> modified: src/pmdas/mysql/pmdamysql.pl
>>> new file: src/pmdas/pdns/.gitignore
>>> new file: src/pmdas/pdns/GNUmakefile
>>> new file: src/pmdas/pdns/Install
>>> new file: src/pmdas/pdns/Remove
>>> new file: src/pmdas/pdns/pmdapdns.pl
>>> modified: src/pmlogger/pmlogger.c
>>> modified: src/pmns/stdpmid.pcp
>>>
>>> Signed-off-by: Mark Goodwin <mgoodwin at redhat.com>
>>>
>>> commit ef9852c99666359cdacb1afe65dcac16460ea0ab
>>> Author: Mark Goodwin <mgoodwin at redhat.com>
>>> Date: Thu Aug 6 11:05:12 2009 +1000
>>>
>>> Changes to the Fedora RPM spec in response to review feedback:
>>> - Incorporate a large number of suggestions and cleanups from Eric Sandeen
>>> - Greatly simplified the %post and %postun scriptlets - remove ancient crud
>>> - Added ldconfig scriptlets for -libs
>>> - Don't explicitly require IB libs, since they're libs and RPM figures it out
>>> - No need to explicitly BuildRequire gcc-c++ libstdc++-devel
>>> - Add BuildRequires on perl-ExtUtils-MakeMaker
>>> - Remove explicit ia64 Requires: libunwind
>>> - Have pcp-libs require the base package since Fedora insists
>>> - Add and clarify some comments
>>> - Create %{_localstatedir}/run/pcp and ship it (so it'll be removed)
>>> - Use %doc for CHANGELOG COPYING INSTALL README VERSION.pcp pcp.lsm
>>>
>>> Signed-off-by: Mark Goodwin <mgoodwin at redhat.com>
>>>
>>> commit 88c3b113c45f61d8959b95f0486d5d625b8c85af
>>> Author: Mark Goodwin <mgoodwin at redhat.com>
>>> Date: Thu Aug 6 10:17:51 2009 +1000
>>>
>>> git ignore generated gram*.tab.c
>>>
>>> Signed-off-by: Mark Goodwin <mgoodwin at redhat.com>
>>>
>>> commit 0fe594271c4635432616eaec4640b38773d150cc
>>> Author: Mark Goodwin <mgoodwin at redhat.com>
>>> Date: Thu Aug 6 10:12:55 2009 +1000
>>>
>>> preserve generated gram.tab.c because debuginfo needs it
>>>
>>> Signed-off-by: Mark Goodwin <mgoodwin at redhat.com>
>>>
>>> commit 385ce8e19c550792fd61a4e6112408c208b75649
>>> Author: Mark Goodwin <mgoodwin at redhat.com>
>>> Date: Thu Aug 6 09:21:48 2009 +1000
>>>
>>> fix minor typo in comment
>>>
>>> Signed-off-by: Mark Goodwin <mgoodwin at redhat.com>
>>>
>>> commit b8cb2dd7c9b5d33e473bf0d89a64da3c2b2eec92
>>> Merge: 1c47a24... 684232a...
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Wed Aug 5 16:52:11 2009 +1000
>>>
>>> Merge branch 'solaris' of git://oss.sgi.com/makc/pcp into dev
>>>
>>> commit 1c47a24dbef502d09439e663b727a2bf85e5499d
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Wed Aug 5 16:51:47 2009 +1000
>>>
>>> Ignore all MMV files that begin with a period.
>>>
>>> commit 4bc076258e025f86b6215783933f7adcb77ace90
>>> Author: Max Matveev <makc at gmx.co.uk>
>>> Date: Wed Aug 5 02:10:37 2009 +1000
>>>
>>> Add network metrics to Solaris PMDA
>>>
>>> Use MIB2 interface available via TI to extract IPv4 interfaces and UDP stats
>>> from the kernel.
>>>
>>> The good/bad fetch is on the whole group - if any of the sources fail (even if
>>> we don't use them like IPv6 interface stats) then whole fetch will return no
>>> data.
>>>
>>> Fetch callback still uses offset for the metrics which have indoms, for
>>> singular metrics a pointer to the global storage is used whence the change
>>> of type in md_offset.
>>>
>>> commit 4e1e119bfb7cfa6d09c81049fec95d3dc9e13356
>>> Author: Mark Goodwin <mgoodwin at redhat.com>
>>> Date: Tue Aug 4 12:19:07 2009 +1000
>>>
>>> dont install trace demo binaries since src is installed anyway
>>>
>>> Signed-off-by: Mark Goodwin <mgoodwin at redhat.com>
>>>
>>> commit 234ffe3f3badbf73768f114fa692eac0dd7ec079
>>> Author: Max Matveev <makc at gmx.co.uk>
>>> Date: Mon Aug 3 10:33:41 2009 +1000
>>>
>>> Report ZFS statistics from Solaris PMDA
>>>
>>> Use libzfs interfaces to extract some data about ZFS "dataset". Datasets
>>> are filesystems, snapshots and volumes, the reported data include information
>>> about used and available space, compression ratios, number of copies,
>>> reserved space per dataset. Space utilization is additionally broken down
>>> into space used by children, space used by snapshot and space used by the
>>> dataset itself.
>>>
>>> commit 5da67384ae7719254afd7d76d8b10558b9d3de07
>>> Author: Max Matveev <makc at gmx.co.uk>
>>> Date: Mon Aug 3 01:59:01 2009 +1000
>>>
>>> Export zpool stats from Solaris PMDA
>>>
>>> Use libzfs to extract zpool stats like capacity, iops, bandwidth etc
>>>
>>> commit 684232a1b6178de4a07edbd0e3308f49789992fb
>>> Author: Max Matveev <makc at gmx.co.uk>
>>> Date: Sun Aug 2 23:05:11 2009 +1000
>>>
>>> Put 64 bit libraries into the right place on Solaris
>>>
>>> Solaris keeps its 64 bit shared libraries in /usr/lib/64 - use the pointer
>>> size to decide if we're building 32 or 64 bit version and push the library
>>> in the right place.
>>>
>>> commit 3cf745f62c0099577265994de97c0a755cb06f9b
>>> Author: Max Matveev <makc at gmx.co.uk>
>>> Date: Sun Aug 2 04:11:32 2009 +1000
>>>
>>> Generate packages for Solaris
>>>
>>> commit 74c7543c88bc3776b81ae8a5eab4de940537ffdd
>>> Author: Max Matveev <makc at gmx.co.uk>
>>> Date: Sun Aug 2 03:03:35 2009 +1000
>>>
>>> Allow user to choose her compiler
>>>
>>> New autoconf attempts to compile the headers as part of AC_HEADER_CHECK.
>>> It AC_PROG_CC is not called then AC_HEADER_CHECK generates scary warning -
>>> just call AC_PROG_CC unconditionally: it will honour CC from the environment
>>> but it will also do the right things about later checks.
>>>
>>> On Solaris if the alternative compiler is not GCC then don't feed it
>>> GNU options.
>>>
>>> commit 5dc76a384254b3a0b1a72ff1be5faec042effa73
>>> Author: Mark Goodwin <mgoodwin at redhat.com>
>>> Date: Fri Jul 31 15:44:40 2009 +1000
>>>
>>> RPM spec specifically for the Fedora Project.
>>>
>>> Signed-off-by: Mark Goodwin <mgoodwin at redhat.com>
>>>
>>> commit 1274605472853400f4297439c433a05829735a64
>>> Author: Mark Goodwin <mgoodwin at redhat.com>
>>> Date: Fri Jul 31 15:40:58 2009 +1000
>>>
>>> tweak configure to move PCP_BINADM_DIR out of /usr/share, into /usr/lib.
>>> Arch dependent binaries should not be installed below /usr/share.
>>>
>>> Signed-off-by: Mark Goodwin <mgoodwin at redhat.com>
>>>
>>> commit 17700e2bdaa0060178c3d025edb92280f36451f6
>>> Author: Mark Goodwin <mgoodwin at redhat.com>
>>> Date: Fri Jul 31 11:47:39 2009 +1000
>>>
>>> delete unneeded explicit script interpreter to keep rpmlint happy
>>>
>>> Signed-off-by: Mark Goodwin <mgoodwin at redhat.com>
>>>
>>> commit b971772c281a5aff6408981e7086cbfa4bcee6ba
>>> Author: Mark Goodwin <mgoodwin at redhat.com>
>>> Date: Fri Jul 31 11:33:20 2009 +1000
>>>
>>> pmpost does not really need to be setuid
>>>
>>> Signed-off-by: Mark Goodwin <mgoodwin at redhat.com>
>>>
>>> commit 0a4cd4b8bfe44265bdccfb93db5e19c28c796f79
>>> Author: Mark Goodwin <mgoodwin at redhat.com>
>>> Date: Fri Jul 31 11:11:48 2009 +1000
>>>
>>> Nuke migrate_pcp_var_dir, no longer needed
>>>
>>> Signed-off-by: Mark Goodwin <mgoodwin at redhat.com>
>>>
>>> commit bf98168e8a7cea7697497f4972dd835fa10e29b6
>>> Author: Mark Goodwin <mgoodwin at redhat.com>
>>> Date: Thu Jul 30 18:25:12 2009 +1000
>>>
>>> default chkconfig off for all PCP services
>>>
>>> Signed-off-by: Mark Goodwin <mgoodwin at redhat.com>
>>>
>>> commit 679b9d50fa63430066fbf8343607205dd46d20f7
>>> Author: Ken McDonell <kenj at internode.on.net>
>>> Date: Wed Jul 29 20:38:59 2009 +1000
>>>
>>> Initial work for PMDA_INTERFACE_4 and new sample.secret.* dynamic metrics.
>>>
>>> commit 6c17c674360b6d6738f37bb9330fdb607dd1c872
>>> Author: Ken McDonell <kenj at internode.on.net>
>>> Date: Wed Jul 29 20:38:07 2009 +1000
>>>
>>> Remove log files that were added by mistake.
>>>
>>> commit a707483a5af651b430b9e2b4a193a5b64585e6e9
>>> Author: Ken McDonell <kenj at internode.on.net>
>>> Date: Wed Jul 29 20:34:10 2009 +1000
>>>
>>> Second round of PMDA_INTERFACE_4 and dynamic PMNS changes.
>>>
>>> Gets things working for DSO PMDAs.
>>>
>>> commit 8e0f63dbb6ae77a2d526e9472b85d8de1b5bab84
>>> Author: Ken McDonell <kenj at internode.on.net>
>>> Date: Wed Jul 29 20:31:52 2009 +1000
>>>
>>> Extending domain number to 9 bits (continued).
>>>
>>> commit f2866c9e1bf8c9a70213ed4cd859d437b9af6bad
>>> Author: Mark Goodwin <goodwinos at gmail.com>
>>> Date: Mon Jul 27 19:59:39 2009 +1000
>>>
>>> Bump version to 2.9.0.
>>>
>>> commit d3f630a269c73017407b46713bb75b91a7516475
>>> Author: Mark Goodwin <goodwinos at gmail.com>
>>> Date: Mon Jul 27 19:57:22 2009 +1000
>>>
>>> Tweak Debian perl packaging with explicit dependencies, resolving build warnings.
>>>
>>> commit e69333488d42d8b3b1a8f35ea1f9ea3b2f12ad35
>>> Author: Mark Goodwin <goodwinos at gmail.com>
>>> Date: Mon Jul 27 19:56:41 2009 +1000
>>>
>>> Make the pcp2.x RPM provide pcp-devel and pcp-libs
>>>
>>> This RPM packaging change makes it easier for dependent
>>> packages such as pcp-gui to transition to pcp-3.x, where
>>> we have split packaging: pcp, pcp-devel and pcp-libs.
>>>
>>> commit 3687851ac4ee2ccc1810b25fccb31c9935427cdf
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Mon Jul 27 15:23:21 2009 +1000
>>>
>>> Split apart the POSIX and Win32 daemon PMDA code in pmcd.
>>>
>>> No longer use a wrapper for the Win32 implementation either, which
>>> allows us to have finer control over opened/closed file descriptors
>>> and gives us a better shot at pipe agents (still not functional due
>>> to fd inheritance, I think) and also keeps track of the child PIDs.
>>>
>>> commit ee52cf1d530317ed290e07bf3dc964b362e3ca5c
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Mon Jul 27 14:07:54 2009 +1000
>>>
>>> When starting a PMDA in socket mode, we must mark the file descriptor as such.
>>>
>>> In order for socket communication to function on Windows, socket fd's must be
>>> flagged (and hence distinct from pipes, so we know whether to write/send there).
>>> This gets socket-based daemon PMDAs to startup correctly under Win32 conditions.
>>>
>>> commit bd5862657c7f532369ea6e760a06a0aa9f81f74f
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Mon Jul 27 14:04:17 2009 +1000
>>>
>>> Fix a typo in a comment in pmcd code.
>>>
>>> commit 61c43f8885220a6e7d41624ace23d1157ef90e29
>>> Merge: 910d640... 3baf8e7...
>>> Author: Mark Goodwin <mgoodwin at redhat.com>
>>> Date: Mon Jul 27 09:37:33 2009 +1000
>>>
>>> Merge branch '3.0.0' of git+ssh://markgw@oss.sgi.com/oss/git/markgw/pcp/pcp into dev
>>>
>>> commit 3baf8e7a6491c9533a1b0ff00c6d711a32677536
>>> Author: Mark Goodwin <mgoodwin at redhat.com>
>>> Date: Mon Jul 27 09:32:57 2009 +1000
>>>
>>> take -fstack-protector-all back out again since it's only supported with newer compilers. Probably should add a configure test for it at some stage
>>>
>>> Signed-off-by: Mark Goodwin <mgoodwin at redhat.com>
>>>
>>> commit eb409a0cb7c638a42e2546d7f6e67ac1aa0ac66c
>>> Author: Josef 'Jeff' Sipek <jeffpc at josefsipek.net>
>>> Date: Mon Jul 27 08:45:50 2009 +1000
>>>
>>> PowerDNS PMDA's packetcache_size and qsize_q are instant values, not counters
>>>
>>> commit c8728d91900e220159bc925084cb3a862a4a5d26
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Mon Jul 27 08:44:21 2009 +1000
>>>
>>> Update pmlogger -x message - kill is now (cross-platform) pmsignal.
>>>
>>> commit 47a82efaefdc26f37ce62df6d736092b96c2ae61
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Thu Jul 23 15:18:20 2009 +1000
>>>
>>> Tweak deb package long descriptions for bug 537563.
>>>
>>> commit 0b9600ede2f03f15d271ab812a703d0b606c4506
>>> Author: Josef 'Jeff' Sipek <jeffpc at josefsipek.net>
>>> Date: Thu Jul 23 14:49:30 2009 +1000
>>>
>>> Add a new PMDA for the PowerDNS daemon (www.powerdns.com).
>>>
>>> commit df6f93cd1b2de0fc1e531cc568f4e53b73adf7e5
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Thu Jul 23 14:32:59 2009 +1000
>>>
>>> Remove a duplicated metric from the MySQL PMDA.
>>>
>>> Problem reported by Josef Sipek, install of this PMDA now
>>> fails due to reworking of the pmns generation code, before
>>> this would have silently overwritten one metric with the
>>> other (122 won in our production environment so I've kept
>>> that) but now we check for this condition.
>>>
>>> commit 737b6904574678e174e5d8b969b91edce66c759a
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Thu Jul 23 13:26:31 2009 +1000
>>>
>>> Improve error reporting in pmcd on Windows.
>>>
>>> Should we be asked to open a UNIX domain socket channel to a
>>> PMDA on Windows, improve the message for easier diagnosis.
>>>
>>> commit c87287e790d4636e2abbd98f223daa0c67c644e8
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Thu Jul 23 13:25:11 2009 +1000
>>>
>>> Improve error reporting in PCP::PMDA perl module.
>>>
>>> When failing to build the PMNS file, additional diagnostics
>>> would be helpful to pinpoint the source of the error.
>>>
>>> commit 910d640b1cdd749de83245523a24e2486a8f5cc2
>>> Merge: 7f3bf2c... 0c1b4b2...
>>> Author: Mark Goodwin <mgoodwin at redhat.com>
>>> Date: Thu Jul 23 11:39:22 2009 +1000
>>>
>>> Merge branch '3.0.0' of git://oss.sgi.com/markgw/pcp/pcp into dev
>>>
>>> commit 0c1b4b2f4bd9b0759964fb056a1469b2dd052ce6
>>> Author: Mark Goodwin <mgoodwin at redhat.com>
>>> Date: Wed Jul 22 17:02:40 2009 +1000
>>>
>>> Tweak RPM dependency rules for 2.x -> 3.x upgrades and
>>> massage the PCP RPM spec to reduce rpmlint noise.
>>>
>>> Also added -fstack-protector-all to Linux PCFLAGS.
>>>
>>> On branch 3.0.0:
>>> modified: build/rpm/pcp.spec.in
>>> modified: src/include/builddefs.in
>>>
>>> Signed-off-by: Mark Goodwin <mgoodwin at redhat.com>
>>>
>>> commit 7f3bf2c9a2966801f95c31f049291da8434faae4
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Wed Jul 22 13:37:47 2009 +1000
>>>
>>> Ensure the Perl PMDA process runs in its own process group,
>>> and ignore SIGTERM in the atexit handler. This resolves an
>>> issue with the earlier fix for reaping children.
>>>
>>> commit d597f63b47a9eb8140c878bbc26c1434431c3096
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Wed Jul 22 11:52:09 2009 +1000
>>>
>>> Comment out some diagnostics in named PMDA.
>>>
>>> commit c9b822c236f1bf219d830d1185d16cc9e87eae4c
>>> Author: Rodney McKee <rmckee at aconex.com>
>>> Date: Wed Jul 22 11:44:32 2009 +1000
>>>
>>> Add the named PMDA, initially supporting BIND 9.4.
>>>
>>> commit cf003f4dddc1f26d2d7818f7c17bfd942a448ea2
>>> Merge: fee86f6... fcca602...
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Wed Jul 22 10:18:55 2009 +1000
>>>
>>> Merge branch 'dev' of git://oss.sgi.com/mort/pcp into dev
>>>
>>> commit fcca6023cd689ea5f48c6f6c538e2e21af67d32e
>>> Author: Martin Hicks <mort at socrates.bork.org>
>>> Date: Tue Jul 21 17:34:05 2009 -0400
>>>
>>> Do not create infiniband pmda config file by default
>>>
>>> In order to simplify monitoring in the simple case where a
>>> machine has a single Infiniband interface, the config file is
>>> no longer created by default. Instead, only local ports are
>>> monitored. This does solve the problem where if an infiniband
>>> card is replaced in a machine that the administrator no longer
>>> has to remove the /var/lib/pcp/pmdas/ib/config file in order
>>> to restore infiniband monitoring.
>>>
>>> For more advanced usage where static instance naming is required
>>> or remote Infiniband ports must be monitored a config file can
>>> still be created, but it is the responsibility of the administrator
>>> to update the GUIDs if an HCA replacement is done.
>>>
>>> commit 840243974c106daae11368730043534d4684bc59
>>> Author: Martin Hicks <mort at socrates.bork.org>
>>> Date: Tue Jul 21 17:31:54 2009 -0400
>>>
>>> Manpage for the infiniband PMDA
>>>
>>> commit fee86f6503252207a236c5eb81c8d092eede1a87
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Thu Jul 16 16:59:45 2009 +1000
>>>
>>> Rework the special SGI CPU numbering logic in the Linux PMDA.
>>>
>>> This resolves a buffer overflow bug in the use of realpath(3)
>>> in this routine, which was being passed a too-small buffer.
>>> This issue was initially reported to us by Chandana De Silva,
>>> but looks like others have come across it too, e.g.:
>>> https://bugs.launchpad.net/ubuntu/+source/pcp/+bug/384776
>>>
>>> I've removed the unused "cname" field from the per-CPU data
>>> structure, and also added some short-circuit logic into this
>>> code so that we bail out of this logic early if no hardware-
>>> graph filesystem paths are detected (which will be the normal
>>> case for most people).
>>>
>>> commit 67ae7fd90a32a43bbc97b939e57302556e864459
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Thu Jul 16 16:54:11 2009 +1000
>>>
>>> Resolve a valgrind reported problem when initing Linux loadavg metric.
>>>
>>> Ensure read buffer is initialised to zero, and also check the return
>>> code for a failed read syscall.
>>>
>>> commit 7f0167d8410147a6645699b0a385dc0c3bdfa5ac
>>> Author: Martin Hicks <mort at bork.org>
>>> Date: Tue Jul 14 14:14:31 2009 -0400
>>>
>>> Reserve domain ID 99 for SGI Ultra Violet PMDA
>>>
>>> commit de5188e9814f25c043dd278d3b3b9be0f3d8a306
>>> Author: Ken McDonell <kenj at internode.on.net>
>>> Date: Sun Jul 12 07:51:13 2009 +1000
>>>
>>> Increase domain from 8-bits to 9-bits.
>>>
>>> Cleaned up the use of the top bit as a flag for both a pmID and
>>> a pmInDom ... now using a 1 bit flag field instead of the 2-bit
>>> "pad" field. Changes required for PMNS parsing in libpcp and
>>> internal handling of help text.
>>>
>>> Introduced PMDA_INTERFACE_4, but no code using this yet.
>>>
>>> Encoding and logic to handle dynamic PMNS entries introduced and
>>> some local support in parsing and reporting local PMNS entries
>>> (syntax for PMID in PMNS is <domain>:*:*) ... this means domain
>>> number 511 (DYNAMIC_PMID) is now special and reserved.
>>>
>>> commit 3a0217c80b84c56d21bdd64f9d6396c9c583556d
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Fri Jul 10 10:37:08 2009 +1000
>>>
>>> Add a PCP::PMDA module check into PMDA Install path for Perl PMDAs.
>>>
>>> commit 2c278354dedbb229721ffbb6cd1934896095c4a7
>>> Author: Nathan Scott <nathans at debian.org>
>>> Date: Fri Jul 10 10:35:08 2009 +1000
>>>
>>> Make Perl PMDA module explicit wrt libpcp_pmda interface version used.
>>>
>>>
>>>
>>> _______________________________________________
>>> pcp mailing list
>>> pcp at oss.sgi.com
>>> http://oss.sgi.com/mailman/listinfo/pcp
>
More information about the pcp
mailing list