Send inn-workers mailing list submissions to
[email protected]
To subscribe or unsubscribe via the World Wide Web, visit
https://lists.isc.org/mailman/listinfo/inn-workers
or, via email, send a message with subject or body 'help' to
[email protected]
You can reach the person managing the list at
[email protected]
When replying, please edit your Subject line so it is more specific
than "Re: Contents of inn-workers digest..."
Today's Topics:
1. Re: INN 2.5.4 strange crash (Julien ?LIE)
2. Re: INN 2.5.4 strange crash (Petr Novopashenniy)
----------------------------------------------------------------------
Message: 1
Date: Tue, 03 Feb 2015 14:00:47 +0100
From: Julien ?LIE <[email protected]>
To: [email protected]
Subject: Re: INN 2.5.4 strange crash
Message-ID: <[email protected]>
Content-Type: text/plain; charset=US-ASCII; format=flowed
Hi Petr,
>> However, I am still puzzled as for the origin of the error.
>> innfeed-neva2! is removed from sleeping (SCHANremove just before that
>> log)
>> but is still considered to be sleeping the next run of the loop (and
>> again
>> without Waker...).
>
> I saw "blocked sleeping" before (for my two channels, innfeed-neva1 and
> innfeed-neva2 at varios times), but not in the time of crash.
They do not seem related then.
I do not see what a proper fix would be for that annoying crash.
As the state of the channel is not the one it should be, and appears to
be
unusable even after removing its sleeping state, I suggest to just close
it.
Could you please change in chan.c:
if (cp->Waker != NULL) {
(*cp->Waker)(cp);
} else {
warn("%s %d sleeping without Waker",
CHANname(cp), fd);
}
to:
if (cp->Waker != NULL) {
(*cp->Waker)(cp);
} else {
name = CHANname(cp);
warn("%s %d sleeping without Waker", name, fd);
SITEchanclose(cp);
CHANclose(cp, name);
}
and tell us if it works better?
I hope you will see "%s %d sleeping without Waker" only once, then the
channel
closes properly and reopens fine a bit later.
When you have such logs, please paste them so as to check the closing is
fine.
--
Julien
------------------------------
Message: 2
Date: Tue, 3 Feb 2015 16:21:17 +0300 (MSK)
From: Petr Novopashenniy <[email protected]>
To: Julien ?LIE <[email protected]>
Cc: [email protected]
Subject: Re: INN 2.5.4 strange crash
Message-ID: <[email protected]>
Content-Type: TEXT/PLAIN; charset=US-ASCII
On Tue, 3 Feb 2015, Julien ?LIE wrote:
J?? Hi Petr,
J??
J?? > > However, I am still puzzled as for the origin of the error.
J?? > > innfeed-neva2! is removed from sleeping (SCHANremove just before that
J?? > > log)
J?? > > but is still considered to be sleeping the next run of the loop (and
J?? > > again
J?? > > without Waker...).
J?? >
J?? > I saw "blocked sleeping" before (for my two channels, innfeed-neva1 and
J?? > innfeed-neva2 at varios times), but not in the time of crash.
J??
J?? They do not seem related then.
J??
J??
J?? I do not see what a proper fix would be for that annoying crash.
J?? As the state of the channel is not the one it should be, and appears to be
J?? unusable even after removing its sleeping state, I suggest to just close it.
J??
J?? Could you please change in chan.c:
J??
J?? if (cp->Waker != NULL) {
J?? (*cp->Waker)(cp);
J?? } else {
J?? warn("%s %d sleeping without Waker", CHANname(cp),
J?? fd);
J?? }
J??
J?? to:
J??
J?? if (cp->Waker != NULL) {
J?? (*cp->Waker)(cp);
J?? } else {
J?? name = CHANname(cp);
J?? warn("%s %d sleeping without Waker", name, fd);
J?? SITEchanclose(cp);
J?? CHANclose(cp, name);
J?? }
J??
J?? and tell us if it works better?
J?? I hope you will see "%s %d sleeping without Waker" only once, then the
J?? channel
J?? closes properly and reopens fine a bit later.
J??
J?? When you have such logs, please paste them so as to check the closing is
J?? fine.
Ok, I apply this patch, and waiting results now..
Thanks, Julien!
--pety
------------------------------
_______________________________________________
inn-workers mailing list
[email protected]
https://lists.isc.org/mailman/listinfo/inn-workers
End of inn-workers Digest, Vol 69, Issue 2
******************************************
