problem with RECOVERED

[Dominique Frise]

Hi,

We are using 4.2-alfa-20060423 (snapshot of 25th may) and see problems with 
recovered messages/SMS: they sometimes don't get sent.

We have a general Macro in hobbit-alerts.cfg that is used by all mail or SMS rules:

##### Mail & SMS defaults ####
$MAIL_DEFS=REPEAT=24h RECOVERED
$SMS_DEFS=COLOR=red FORMAT=SMS $MAIL_DEFS

##### Mail & SMS recipients ####
$MAIL_DF=MAIL example at example.com $MAIL_DEFS
$SMS_DF=SCRIPT $SENDSMS 0123456789 $SMS_DEFS

#### Example rule ####
HOST=argos,bigmac,docusurf,eliot,fmp7web,ged,hippocrate,iris,mailc,midas,zeb1,zeb2,zelda,zelda2
$MAIL_DF
$SMS_DF


In the following notifications.log you can see that recovered message never got 
sent for hobbitd:

Thu May  4 05:01:31 2006 uldns2.smtp (130.223.4.5) abc.def at xy.ch 1146711691 725
Thu May  4 05:01:46 2006 uldns2.smtp (130.223.4.5) abc.def at xy.ch 1146711706 725 15
Thu May  4 06:03:57 2006 iris.hobbitd (130.223.27.23) abc.def at xy.ch 1146715437 0
Thu May  4 07:02:24 2006 falco.sslcert (130.223.5.21) www.ddd at xy.ch 1146718944 0
Thu May  4 07:05:54 2006 mailc1.msgs (130.223.27.61) root at localhost 1146719154 400
Thu May  4 07:10:55 2006 mailc1.msgs (130.223.27.61) root at localhost 1146719454 
400 300


The history log shows that hobbitd was recovered after 2 cycles (10 min. 1 sec).

Last 50 log entries (Full HTML log)
Date 				Status 	Duration
Thu May 04 06:13:58 2006 	green 	1:38:44
Thu May 04 06:03:57 2006 	yellow 	0:10:01


I have another example for a procs test where the duration was 23 min.

BTW: a man page for log formats would be helpful.



Dominique
UNIL - University of Lausanne