From mboxrd@z Thu Jan 1 00:00:00 1970 From: Fabio M. Di Nitto Date: Mon, 14 Jul 2008 19:45:30 +0200 (CEST) Subject: [Cluster-devel] Cluster Project branch, master, updated. cluster-2.99.05-66-g0dcc5e5 In-Reply-To: <20080714155251.GA11644@redhat.com> References: <20080711171314.14780.qmail@sourceware.org> <20080714155251.GA11644@redhat.com> Message-ID: List-Id: To: cluster-devel.redhat.com MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit On Mon, 14 Jul 2008, David Teigland wrote: > On Mon, Jul 14, 2008 at 09:57:37AM +0200, Fabio M. Di Nitto wrote: >>> +#define DEFAULT_MODE LOG_MODE_OUTPUT_SYSLOG_THREADED | \ >>> + LOG_MODE_NOSUBSYS | \ >> >> isn't LOG_MODE_NOSUBSYS going to remove [FENCED] from the output? >> >> How are we going to recognize our messages in syslog? >> >> My understading was that NOSUBSYS should be used to print to stderr only > > I added the flag after noticing that syslog already prefixes it, e.g. > > Jul 11 13:30:56 bull-01 fenced[4214]: [fenced] cluster is down, exiting > > But the real reason we'd want it, I suspect, is if someone sends all > cluster errors to a common /var/log/cluster/cluster.log, in which case I > think we'd want logsys to add a prefix... > Maybe the NOSUBSYS implementation in logsys needs a change. I think we always want SUBSYS when writing to file but we can skip it when writing to syslog and stderr. Tho I am not sure it's actually doable because the string format is indipendent from the target log. Lon should know this bit better than me since he implemented it :) Fabio -- I'm going to make him an offer he can't refuse.