Funny logging of console messages in /usr/adm/SYSLOG

David Hinds dhinds at portia.Stanford.EDU
Sat Nov 3 12:04:41 AEST 1990


Someone else posted a description of something similar to this a few days
ago, and I think we have the same problem.  It seems that when "grcond"
starts up, among the various messages that get put in SYSLOG is a kind of
history of all sorts of old console error messages that have accumulated
over a period of weeks.  In looking through our SYSLOG, I found the same
group of messages - including a few disk-full errors, a bunch of SCSI tape
errors, and some other harmless things - repeated dozens of times.  Each
time a new message was logged (distinguishable because they were logged
by "unix:"), it was added to the usual grcond complaint list.  This list
is up to 20 messages now, not counting repeats.  It seems to defeat the
purpose of SYSLOG to have hundreds of bugus messages accumulating all the
time - is there anyway to tell grcond to shut up?

(This isn't SO bad - before I figured out how to turn off timeslave's
default logging, we were getting 250K of time messages per week)

 -David Hinds
  dhinds at cb-iris.stanford.edu



More information about the Comp.sys.sgi mailing list