Anope IRC Services
Denora Stats (discontinued) => Denora 1.5.x Support (Read Only) => Topic started by: CrazyCat on May 19, 2010, 09:21:49 AM
-
Hello there,
I've noticed a trouble for months, I can't understand it.
When I launch denora (/home/ircd/stats/denorarc start or /home/ircd/stats/stats), it's all right. But sometimes, the stats.pid seems to be erased, and the denora.chk makes denora trying to restart, I can see its connections each minute.
Here is my crontab:
* * * * * /home/ircd/stats/denora.chk > /dev/null 2>&1
And denora.chk is the example.chk renamed, it works well.
So, what could erase the stats.pid ? Is there a way to log what happen ? I had this trouble with a previous server, I've changed it (with new install of OS, no particular process which can erase this file) a few days ago and the same thing happens yesterday.
Thanks
-
yeah i can vaguely remember having the same problem, happened rarely but sometimes happened. even with anope. might check with the anope team if they are aware of such an issue and if it was fixed for them.
-
I don't think it was fixed in anope : since a few days, I've the trouble with anope 1.8.4.
And it's really strange, I can't find any logic in the deletion, seems to be a random event.
-
I've not heard anyone mention this in anope for a long time....
-
Actually we have this problem quite often and fairly recently as well with 1.9.1p1. I always attributed it to the cron script which doesn't seem to work sufficiently most of the time. Since we made our own generic cron script the problems seem to have resolved themselves.
-
Worth using yours instead of ours? :)
(otherwise I'll look at ours later on)
-
Worth using yours instead of ours? :)
(otherwise I'll look at ours later on)
Actually "ours" has always done this. At least since 1.6. The script deletes the pid for some unknown to me reason. IIRC, the system will create/overwrite the .pid file to start with. I think back in about 07/08 I asked GD to explain why "our" script deletes the pid and while the explanation he gave me made sense, I cant exactly remember why or why it only happens sometimes. But this is an old issue and is/was in the bugtracker at one point. Personally I just used a modified version of Unreal's ircdchk script.
-
Actually "ours" has always done this.
Nobody said it didn't.
-
In the cronscript packed with anope it seems the $ANOPIDF is used for the pid file, but later on in the if where it greps that it uses $ANOPID Missed a letter :p I don't know anything about shell scripting really, but is this the problem perhaps?
-
Nobody said it didn't.
Just like no one claimed that you was saying it never happened before. I'm sorry, I assumed that you'd be able to understand what I was referring to as long as you have known me. I'll be more clear in the future.
-
In the cronscript packed with anope it seems the $ANOPIDF is used for the pid file, but later on in the if where it greps that it uses $ANOPID Missed a letter :p I don't know anything about shell scripting really, but is this the problem perhaps?
No, the ANOPIDF is used to set the ANOPID variable (with a cat), and it's ANOPID (the PID) which must be used in the grep.
And if the .chk was faulty, I guess it may be always faulty, not only sometimes.
-
That makes no sense to me, where does it get $ANOPID from? It's not set anywhere lol
Nevermind... I'm blind :| but anyways, this does happen to mine not as much on this new shell of mine, maybe its a crontab problem? the software i mean
-
The main part of the services.chk is:
if [ -f $ANOPIDF ]
then
ANOPID=`cat $ANOPIDF`
if [ `ps auwx | grep $ANOPROG | grep $ANOPID | grep -v -c grep` = 1 ]
then
exit
fi
rm -f $ANOPIDF
fi
Third line: ANOPID=`cat $ANOPIDF`. It means that ANOPID is set with the content of ANOPIDF (services.pid)
Edit: whoops, I answered you when you were editing your message :)
I don't think it's crontab himself, but peharps a side effect if the server is overload, and the ps auwx don't respond fast enough ?