infxalrm?

--Drove_of_Sheep_030_000 Content-Type: TEXT/plain; charset=us-ascii Content-MD5: j+i36o5oDaPTLkJxGmkZCA== Listers, I've checked the faq, the unofficial faq, and the TechInfo center's "Product Defect Query" for references to infxalrm. I've found no references thus far. The engineer I'm working with from tech support is insisting that I reboot my high availability server and install the default ALARMPROGRAM. I will do this as a last resort, but I'm hoping someone on the list has encountered what I am seeing. Please read me the riot act if I am missing something obvious. My problem is that I'm finding a large number of "infxalrm" processes on my server(s). The number seems to be proportional to either the number of CPUs or the amount of DB activity. When I say a large number, I'm talking ~180. The processes do not terminate when the instance is bounced. I can kill them: root@marauder:37 /# ps -edalf | grep 4644 | grep -v grep 8 S informix 4644 19168 0 69 10 65de9908 114 65de9978 Oct 16 ? 0:00 /bin/sh -c /home1/informix/bin/infx 8 S root 4649 4644 0 69 10 64c70230 561 62393b8c Oct 16 ? 0:00 /home1/informix/bin/infxalrm 2 23 L root@marauder:38 /# kill -9 4649 root@marauder:39 /# !ps ps -edalf | grep 4644 | grep -v grep root@marauder:40 /# In the online log, I see: 07:36:40 Process exited with return code 137: /bin/sh /bin/sh -c /home1/informix/bin/infxalrm 2 23 "Logical Log 38613Complete." "Logical Log 38613 Complete." I'm not sure where to look to interpret the return code; for Solaris errno.h has: #define ENOTCONN 134 /* Socket is not connected */ /* XENIX has 135 - 142 */ #define ESHUTDOWN 143 /* Can't send after socket shutdown */ When I kill one of these processes that was started before the instance was last bounced, the process goes away but there is no message in the online.log. From what I can gather, infxalrm is triggered by certain DB "events" such as logical log completion, disk errors, etc. It is clear to me that not every event is resulting in one of these hung processes. I'm also seeing the problem with events other than logical log comple- tion. Tech support is saying that the problem lies with the script referenced by the ALARMPROGRAM setting. I've tried several different scripts, but it doesn't seem to matter. I have included the script that I was using when I first found the problem. If I had to hazard a guess as to what's going on, I'd say that these processes are hanging because they couldn't communicate with the engine at some critical point. None of the other apps using the server have connectivity problems. This began when I upgraded from 7.23.UC4 to 7.30.UC3. HARDWARE: Sun UE4000 PRODUCT: Informix Dynamic Server Version 7.30.UC3 OS: Solaris 2.6 TIA. Larry Kemmerling Product Systems Engineer AT&T Wireless Services, Aviation Communications Division --Drove_of_Sheep_030_000 Content-Type: APPLICATION/x-sun-shell-script; name="assert_fail.sh"; x-unix-mode=0644 Content-Transfer-Encoding: BASE64 Content-Description: assert_fail.sh Content-MD5: PamPrv1+BwfZrQheFiXDDg== #! /bin/sh #******************************************************************************* # PROG='Informix_Error_Detect' DATE=`date +%y%m%d%H%M%S` TMPFILE=${TMPDIR:-/tmp}/$PROG.$DATE USER_LIST=larryk@airone.claircom.com,scottm@airone.claircom.com SERVER=`echo $INFORMIXSERVER` EXIT_STATUS=0 EVENT_SEVERITY=$1 EVENT_CLASS=$2 EVENT_MSG="$3" EVENT_ADD_TEXT="$4" EVENT_FILE="$5" rm -f $TMPFILE touch $TMPFILE if [ $1 -gt 3 ]; then echo " " >> $TMPFILE echo "($SERVER) is having problems with the Informix database" >> $TMPFILE echo " " >> $TMPFILE echo " Event_Severity ($EVENT_SEVERITY)" >> $TMPFILE echo " Event_Class ($EVENT_CLASS)" >> $TMPFILE echo " Event_Msg ($EVENT_MSG)" >> $TMPFILE echo " EVENT_ADD_TEXT ($EVENT_ADD_TEXT)" >> $TMPFILE echo " EVENT_FILE ($EVENT_FILE)" >> $TMPFILE echo " " >> $TMPFILE /usr/ucb/Mail -s "$PROG Program Reporting Problems . . . . " $USER_LIST < $TMPFILE fi rm -f $TMPFILE exit $EXIT_STATUS --Drove_of_Sheep_030_000--