[pcp] Reserve pmid used by the CTDB PMDA
ddiss at suse.de
Wed Sep 14 06:00:45 CDT 2011
Thanks for the feedback Ken and Nathan. Sorry for the late response,
I've been out of the office this week.
On Fri, 09 Sep 2011 07:13:18 +1000
Ken McDonell <kenj at internode.on.net> wrote:
> 109 has already been nabbed for the MSSQL PMDA ... if you could move
> to 110 that would help keep the allocation space compact, but if 155
> has already escaped it will be fine to keep that also.
> Just let us know which works for you and we can make it so.
I'll submit a change to upstream ctdb to use 110.
On Mon, 12 Sep 2011 09:10:34 +1000 (EST)
Nathan Scott <nathans at aconex.com> wrote:
> > http://git.samba.org/?p=ctdb.git;a=commit;h=fe7f69d6fc37661c2f1caa11de4ed3a7940d0a2f
> > A domain id of 155 was plucked out of the air for no good reason.
> > I'm willing to change it to 109 (next available) if necessary.
> Hi David,
> The PMDA seems to use a fairly odd PMID numbering scheme (every metric
> seems to define a unique cluster and a unique item number) ... it
> seems like that might bite you as you add more metrics - what was the
> rationale there, OOC?
I used the simple PMDA as a base, which has a somewhat confusing pmns
file. That said, I should have looked at one of the many others.
Here's what I plan to push to ctdb...
The following changes since commit
ReadOnly: revokechild_active is a list, not a context. (2011-09-13
are available in the git repository at:
David Disseldorp (1):
pmda: Use upstream assigned PCP domain id
utils/pmda/domain.h | 2 +-
utils/pmda/pmns | 72
+++++++++++++++++++++++++------------------------- 2 files changed, 37
insertions(+), 37 deletions(-)
More information about the pcp