Changes committed to git://oss.sgi.com/nathans/pcp.git
VERSION.pcp | 2 +-
configure.in | 8 ++++++--
debian/changelog | 6 ++++++
src/pmlogctl/pmlogger_daily.sh | 9 +--------
4 files changed, 14 insertions(+), 11 deletions(-)
commit d2d50468b8cc56cbda6a8e898df74f4180b039a2
Author: Nathan Scott <nathans@xxxxxxxxxx>
Date: Tue Dec 23 14:54:23 2008 +1100
Bump datestamp in version number for RPM upgrades.
commit b6c662ad455ca9e266a93fbe9e6447ebf61834cc
Author: Nathan Scott <nathans@xxxxxxxxxx>
Date: Tue Dec 23 14:52:45 2008 +1100
Resolve an issue with recent versions of autoconf, as reported by
Peter Edwards:
configure.in:292: error: AC_SUBST: `pcp_owner pcp_group' is not a
valid
shell variable name
configure.in:292: the top level
autom4te: /usr/bin/m4 failed with exit status: 1
Looks like we need to be passing only one argument to AC_SUBST,
and that this doubling-up (or 4-at-a-time which we do too) is no
longer valid syntax.
commit 7c8c1d387af7d9f8c82765407d14bec683970d8d
Author: Nathan Scott <nathans@xxxxxxxxxx>
Date: Tue Dec 23 14:48:10 2008 +1100
Correct a logic error in pmlogger_daily scripts detection of running
pmloggers. This recently regressed even more but was actually never
right. Funnily enough, the same logic exists in pmlogger_check and
it _is_ right - so, this just dups that exact logic here. The root
issue is to do with handling of failed glob-expansion on the process
id files below /var/log/pmlogger.
Symptoms of the problem are:
- failure to rotate daily logs for non-primary pmloggers, and
- email messages with diagnostics along the lines:
pmlogger_daily: Error: no pmlogger instance running for host "fw1"
[/var/lib/pcp/config/pmlogger/control:28] ... logging for host "fw1"
unchanged
pmlogger_daily: Error: no pmlogger instance running for host "fw2"
[/var/lib/pcp/config/pmlogger/control:29] ... logging for host "fw2"
unchanged
|