[IRCServices] Complete DB loss on spontaneous SIGHUP

[Ballsy]

FreeBSD 4.2
ircservices-4.5.43
bahamut-1.4.30

	I've had services die for unexplained reasons again twice since 
the original post below...fortunately, the DBs were still intact.  For the 
most recent instance, I had debugging enabled (at level 2 or 3), and I 
discovered the following:

>From my client on the network (which was oper'd):

08:46 [mynetwork]  *** Routing -- from server.az.us.mynetwork.net: Server
          services.mynetwork.net[unknown at 0.0.0.0] closed the connection
08:46 [mynetwork]  services.mynetwork.net was connected for 4339 seconds.  
51/56 sendK/recvK.

And the last log file entries...

[Sep 18 08:45:56.243446 2002] debug: Top of main loop
[Sep 18 08:46:11.253876 2002] debug: Top of main loop
[Sep 18 08:46:11.254066 2002] debug: Checking timeouts at time_msec = 
1561020.214
[Sep 18 08:46:11.254206 2002] debug: Finished timeout list
[Sep 18 08:46:12.937952 2002] Received SIGHUP, restarting.
[Sep 18 08:46:12.938218 2002] debug: Top of main loop
[Sep 18 08:46:12.938343 2002] debug: Running expire routines
[Sep 18 08:46:12.938487 2002] debug: NickServ: updating last seen time for 
`grace
[Sep 18 08:46:12.938586 2002] debug: NickServ: updating last seen time for 
Asellus
[Sep 18 08:46:12.938679 2002] debug: NickServ: updating last seen time for 
Baltazir[LOA]
[Sep 18 08:46:12.938934 2002] debug: NickServ: updating last seen time for 
Ballsy

<snip>  More of the same "updating last seen time" lines

[Sep 18 08:46:12.941303 2002] debug: Saving databases

	
	...and that was it.  It's easy enough to restart services via 
cron, but I'd like to get to the bottom of what's causing this.  In every 
instance, that SIGHUP has been around, but again, I dunno where it's 
coming from.

David



On Tue, 17 Sep 2002, Ballsy wrote:

> 
> FreeBSD 4.2
> ircservices-4.5.43
> 
> 	I migrated from SIRV-2.7.0 (imported DBs using 
> import-db +sirv /my/path/here
> ) 
> fairly successfully today, and services ran fine for most of the day.  I headed 
> home, and when I logged back on, I noticed services were gone.  Logged 
> into the shell and checked the logs, and the last entry was as follows:	
> 
> 
> [Sep 17 17:12:05 2002] Received SIGHUP, restarting.
> 
> 	There was no corresponding RESTART command issued before this, and 
> I can't imagine anyone would have issued a -HUP from the shell since I am 
> the only one that uses it (I'm not naive enough to completely rule it out, 
> of course).  When I restarted the services (unfortunately I didn't check 
> the file sizes of the .db files first), the following were the first 
> entries dumped to the services.log file:
> 
> [Sep 17 17:48:59 2002] Services 4.5.43 (compiled for ircd.dal Bahamut) 
> starting up
> [Sep 17 17:49:00 2002] ChanServ: database load: Deleting founderless 
> channel #7er_himmel
> [Sep 17 17:49:00 2002] ChanServ: database load: Deleting founderless 
> channel #afl
> [Sep 17 17:49:00 2002] ChanServ: database load: Deleting founderless 
> channel #Ai_no_Card
> 
> 	This went on for several lines, as numerous (dare I say all?) 
> channels were deleted.  I have verified that the DBs are being updated 
> every 15m (as per services.conf).  Essentially, apparently due to the 
> seemingly rogue SIGHUP above, my dbs were completely wiped out (even MY 
> nick was showing up as not registered afterwards, and I'd been using it 
> all day).
> 	I searched the archives for a while and couldn't find anything 
> similar, so figured I'd ask here.  Has any seen this type of behaviour 
> before ?  Thanks in advance for any light which can be shed.
> 
> David
> 
> 
> ------------------------------------------------------------------
> To unsubscribe or change your subscription options, visit:
> http://www.ircservices.za.net/mailman/listinfo/ircservices
>