I apologize for the lack of technical detail in this post, but, for the moment I'm just trying to ascertain if I or my syslog-ng config are on drugs. Basically, I have a mixed array of Linux (RedHat 6.2 & 7) and FreeBSD (4.2-S) machines on which I'm running syslog-ng and logging to a central host. (FreeBSD 4.2-S) The problem is that the Linux boxes periodically stop logging. If I send the daemon a HUP, it comes back to life and data starts flowing again, but any log messages that would have appeared while the daemon was hosed are gone. Has anyone experienced anything like this before? Thanks in advance, -Snow
Basically, I have a mixed array of Linux (RedHat 6.2 & 7) and FreeBSD (4.2-S) machines on which I'm running syslog-ng and logging to a central host. (FreeBSD 4.2-S)
The problem is that the Linux boxes periodically stop logging. If I send the daemon a HUP, it comes back to life and data starts flowing again, but any log messages that would have appeared while the daemon was hosed are gone.
Has anyone experienced anything like this before?
what versions of syslog-ng? -sv
On Fri, Apr 06, 2001 at 02:54:28PM -0400, Seth Vidal wrote:
Has anyone experienced anything like this before?
what versions of syslog-ng?
Oops. That would have been a useful bit of information. Sorry. On the Linux machines: syslog-ng 1.4.11, libol 0.2.21 On the BSD machine: syslog-ng 1.4.10, libol 0.2.20 The BSD's are older because I made syslog-ng out of the ports tree. I grabbed the source from the distribution sites when doing the Linux installs. Perhaps I should synchronize the versions? -Snow
On Fri, 6 Apr 2001, James Snow wrote:
I apologize for the lack of technical detail in this post, but, for the moment I'm just trying to ascertain if I or my syslog-ng config are on drugs.
Basically, I have a mixed array of Linux (RedHat 6.2 & 7) and FreeBSD (4.2-S) machines on which I'm running syslog-ng and logging to a central host. (FreeBSD 4.2-S)
The problem is that the Linux boxes periodically stop logging. If I send the daemon a HUP, it comes back to life and data starts flowing again, but any log messages that would have appeared while the daemon was hosed are gone.
Has anyone experienced anything like this before?
Make sure you don't have logrotate setup to run and rotate your syslog files, or if you do make sure it is restarting syslog-ng and not your old syslogd (I got bitten by this myself on one system). The easiest way to check is see if you have a /var/log/messages.1 that is newer than /var/log/messages (or whatever file(s) you are using for your logs). -- Dan Rich <drich@employees.org> | http://www.employees.org/~drich/ | "Danger, you haven't seen the last of me!" | "No, but the first of you turns my stomach!" | -- The Firesign Theatre's Nick Danger
I just fought this issue. If you are running klogd and have syslog-ng configured to likewise log kernel messages (ie.: source kernel_file { file("/proc/kmsg"); };) its most likely hanging upon a read of /proc/kmesg file descriptor. Search the archives a week or two ago, and look for my posts. I document in gory detail what I saw. If you see the same thing, well, then you'll hopefully be able to resolve the issue quickly. On Fri, 6 Apr 2001, James Snow wrote:
I apologize for the lack of technical detail in this post, but, for the moment I'm just trying to ascertain if I or my syslog-ng config are on drugs.
Basically, I have a mixed array of Linux (RedHat 6.2 & 7) and FreeBSD (4.2-S) machines on which I'm running syslog-ng and logging to a central host. (FreeBSD 4.2-S)
The problem is that the Linux boxes periodically stop logging. If I send the daemon a HUP, it comes back to life and data starts flowing again, but any log messages that would have appeared while the daemon was hosed are gone.
Has anyone experienced anything like this before?
Thanks in advance, -Snow
_______________________________________________ syslog-ng maillist - syslog-ng@lists.balabit.hu https://lists.balabit.hu/mailman/listinfo/syslog-ng
Brian Seppanen Charter Communications Regional Data Center 906-228-4226 ext 23 Marquette, MI seppy@chartermi.net
participants (4)
-
Brian E. Seppanen
-
Dan Rich
-
James Snow
-
Seth Vidal