Subject: RE: Problem with "kill -HUP" on syslog-ng - msg#00028
List: syslog-ng
I'm receiving thousands of syslogs per second, so the truss is not that
useful - it's too difficule to isolate the interval. When failed HUP
issue occurrs, no log entry is written to syslog, and the daemon
continues running normally. When a successful HUP occurs, the
appropriate message appears in syslog.
I did not have this issue when I was testing, only when I put the server
under extreme load, so that be a prerequisite for the issue to occur.
Peter
-----Original Message-----
From: syslog-ng-admin@xxxxxxxxxxxxxxxx
[
mailto:syslog-ng-admin@xxxxxxxxxxxxxxxx]On Behalf Of Balazs Scheidler
Sent: Thursday, May 06, 2004 12:59 PM
To: syslog-ng@xxxxxxxxxxxxxxxx
Subject: RE: [syslog-ng]Problem with "kill -HUP" on syslog-ng
2004-05-06, cs keltezéssel 18:39-kor Wyner, Peter ezt írta:
>
I have had similar issues HUPing 1.6.2 under Solaris 2.8. My
resolution
>
is to kill have the script kill -9 the process and start a new one
>
instead of HUPing. The behavior is not consistant - sometimes it will
>
honor the SIGHUP and sometimes it won't.
can you be more specific? it simply does nothing, it crashes or?
syslog-ng should send this message to the system log when it receives a
system log:
notice("SIGHUP received, restarting syslog-ng\n")
do you receive this message when it does not correctly restart? can you
attach truss to the running syslog-ng process when you send a HUP signal
to it?
--
Bazsi
PGP info: KeyID 9AF8D0A9 Fingerprint CD27 CFB0 802C 0944 9CFD 804E C82C
8EB1
_______________________________________________
syslog-ng maillist - syslog-ng@xxxxxxxxxxxxxxxx
https://lists.balabit.hu/mailman/listinfo/syslog-ng
Frequently asked questions at
http://www.campin.net/syslog-ng/faq.html
smime.p7s
Description: S/MIME cryptographic signature
Was this page helpful?
Thread at a glance:
Previous Message by Date:
click to view message preview
RE: Problem with "kill -HUP" on syslog-ng
2004-05-06, cs keltezéssel 18:39-kor Wyner, Peter ezt írta:
> I have had similar issues HUPing 1.6.2 under Solaris 2.8. My resolution
> is to kill have the script kill -9 the process and start a new one
> instead of HUPing. The behavior is not consistant - sometimes it will
> honor the SIGHUP and sometimes it won't.
can you be more specific? it simply does nothing, it crashes or?
syslog-ng should send this message to the system log when it receives a
system log:
notice("SIGHUP received, restarting syslog-ng\n")
do you receive this message when it does not correctly restart? can you
attach truss to the running syslog-ng process when you send a HUP signal
to it?
--
Bazsi
PGP info: KeyID 9AF8D0A9 Fingerprint CD27 CFB0 802C 0944 9CFD 804E C82C 8EB1
_______________________________________________
syslog-ng maillist - syslog-ng@xxxxxxxxxxxxxxxx
https://lists.balabit.hu/mailman/listinfo/syslog-ng
Frequently asked questions at http://www.campin.net/syslog-ng/faq.html
Next Message by Date:
click to view message preview
Tuning for high-throughput
On another topic.....
Anyone have any guidance on optimizing syslog-ng for high-volume
systems? For example, what is a good range for sync(), and are there
any other options that may assist in lowering CPU? We are not using any
variables in the directory of filenames other than $HOST. Here's the
current conf:
options {
log_fifo_size(1024);
dns_cache(yes);
use_dns(yes);
chain_hostnames(no);
keep_hostname(no);
create_dirs(yes);};
source local { sun-streams("/dev/log"); internal(); };
source std_udp { udp( port(52001) ); };
destination all { file("/var/log/messages"); };
destination std { file("/syslog/std/current/$HOST" dir_perm(0755)
perm(0644)); };
log { source(local); destination(all); };
log { source(std_udp); destination(std); };
TIA,
Peter
smime.p7s
Description: S/MIME cryptographic signature
Previous Message by Thread:
click to view message preview
RE: Problem with "kill -HUP" on syslog-ng
2004-05-06, cs keltezéssel 18:39-kor Wyner, Peter ezt írta:
> I have had similar issues HUPing 1.6.2 under Solaris 2.8. My resolution
> is to kill have the script kill -9 the process and start a new one
> instead of HUPing. The behavior is not consistant - sometimes it will
> honor the SIGHUP and sometimes it won't.
can you be more specific? it simply does nothing, it crashes or?
syslog-ng should send this message to the system log when it receives a
system log:
notice("SIGHUP received, restarting syslog-ng\n")
do you receive this message when it does not correctly restart? can you
attach truss to the running syslog-ng process when you send a HUP signal
to it?
--
Bazsi
PGP info: KeyID 9AF8D0A9 Fingerprint CD27 CFB0 802C 0944 9CFD 804E C82C 8EB1
_______________________________________________
syslog-ng maillist - syslog-ng@xxxxxxxxxxxxxxxx
https://lists.balabit.hu/mailman/listinfo/syslog-ng
Frequently asked questions at http://www.campin.net/syslog-ng/faq.html
Next Message by Thread:
click to view message preview
Re: Problem with "kill -HUP" on syslog-ng
On Thu, 6 May 2004 13:04:43 -0400
"Wyner, Peter" <pwyner@xxxxxxxxxxxx> wrote:
> I'm receiving thousands of syslogs per second, so the truss is not that
> useful - it's too difficule to isolate the interval. When failed HUP
> issue occurrs, no log entry is written to syslog, and the daemon
> continues running normally.
This seems to be the same behavior I'm seeing as well. I am not currently
running under light load however so I can test that at the moment.
Running 1.6.2 on a recent Linux 2.4 kernel in a chroot jail, passing -f,
-p, -C and -u options at startup.
John
_______________________________________________
syslog-ng maillist - syslog-ng@xxxxxxxxxxxxxxxx
https://lists.balabit.hu/mailman/listinfo/syslog-ng
Frequently asked questions at http://www.campin.net/syslog-ng/faq.html