I received two replies regarding my question:
Casper Dik <email@example.com> explained that messages are logged
with the string 'syslog:' if the application doesn't do an openlog-call.
Charles Holbrook <firstname.lastname@example.org> suggested checking that enough swapspace
is available and that the number of files in /tmp is not to large. I checked
both of these before posting, lots of swap is available and very few files
Since the message isn't logged by the kernel itself, I started looking at
the applications running on the machine: Our squid-proxy is one of the
large applications on that server. I greped through the sources and voila:
squid is the one logging these messages. We are currently checking different
squid-resources to get a clue what goes wrong. Things would have been
easier if squid had logged the messages with 'squid: '...
Since yesterday, one of our Ultras running 2.5.1 has the following
message in the messages file:
Oct 15 10:17:27 hostname syslog: You've run out of swap file numbers.
The message repeats every couple of hours.
Does anyone know what causes this message? Note that is logged with
facility syslog. I did a 'strings syslogd', but the syslogd-binary
does not contain this string. What else could log this message
with facility syslog? There is plenty of swapspace available.
-- Jan-Olaf Droese | Lahmeyer International GmbH email: email@example.com | Friedberger Straße 173 phone: +49-6101-55-1611 | 61118 Bad Vilbel fax: +49-6101-55-1623 | Germany
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:13:29 CDT