I run a small network of servers running ngircd, three (3) in all. today - about 5 min ago (~20:15 PST) all three ngircd processes abruptly ended. All three of them were ngircd 0.7.7, two were running on Linux and one on FreeBSD. nothing was written to syslog.
...any ideas?
Hi Anthony!
I run a small network of servers running ngircd, three (3) in all. today - about 5 min ago (~20:15 PST) all three ngircd processes abruptly ended. All three of them were ngircd 0.7.7, two were running on Linux and one on FreeBSD. nothing was written to syslog.
Strange.
Did one of the systems generate a "core file"? If yes, it would be useful to have a look at it with gdb to determine where it crashed.
Regards Alex
Nope, none of them did... which just confuses me more. I suppose I should note that I've never experienced this before or since, and I've been using your fine IRCd for many months... It worries me that they all died at once though.
- Anthony
Alexander Barton wrote:
Hi Anthony!
I run a small network of servers running ngircd, three (3) in all. today - about 5 min ago (~20:15 PST) all three ngircd processes abruptly ended. All three of them were ngircd 0.7.7, two were running on Linux and one on FreeBSD. nothing was written to syslog.
Strange.
Did one of the systems generate a "core file"? If yes, it would be useful to have a look at it with gdb to determine where it crashed.
Regards Alex
Am 02.05.2002 um 04:57 schrieb Anthony Kepler:
Nope, none of them did... which just confuses me more.
A lot of systems are configured not to generate core files at all. Have a look at "ulimit".
I suppose I should note that I've never experienced this before or since, and I've been using your fine IRCd for many months... It worries me that they all died at once though.
Right, me too ...
Regards Alex
I keep ulimit set to unlimited... I believe that should be sufficient... right?
Alexander Barton wrote:
Am 02.05.2002 um 04:57 schrieb Anthony Kepler:
Nope, none of them did... which just confuses me more.
A lot of systems are configured not to generate core files at all. Have a look at "ulimit".
I suppose I should note that I've never experienced this before or since, and I've been using your fine IRCd for many months... It worries me that they all died at once though.
Right, me too ...
Regards Alex
Anthony Kepler wrote:
I keep ulimit set to unlimited... I believe that should be sufficient... right?
Right ...
But then, why isn't there a core file at all?? Verry strange ...
Regards Alex
Looking for the simplier explanation... is there a global shutdown command someone who's gained master ops could execute? That would explain all 3 quitting sans core at the same time.
Joshua Coombs
Am 28. Mrz 2004 um 19:26 Uhr schrieb Josh Coombs:
Looking for the simplier explanation... is there a global shutdown command someone who's gained master ops could execute? That would explain all 3 quitting sans core at the same time.
IRC operators can use DIE to shutdown the server they are connected to; but they can't shutdown all servers in a network all at once.
Regards Alex
PS.: Additional note for the archive: you can use REHASH (as IRC operator) to force the server to reload the configuration while running, connected users won't be kicked off. Sometimes quite useful :-)
This is where I apologize to everyone on the list for somehow not noticing that my system clock went back in time to 2002, and has been mislabeling all my outgoing mail. ...I look/feel really smart now.
Alexander Barton wrote:
Am 02.05.2002 um 04:57 schrieb Anthony Kepler:
Nope, none of them did... which just confuses me more.
A lot of systems are configured not to generate core files at all. Have a look at "ulimit".
I suppose I should note that I've never experienced this before or since, and I've been using your fine IRCd for many months... It worries me that they all died at once though.
Right, me too ...
Regards Alex
-
Alexander Barton -
Anthony Kepler -
Josh Coombs