sockstat to rescue.
sockstat -P<mserver procee id>
Helped me realize I had another instance of my app running on another box whose c3p0 max pool size was too high. That was the culprit and consumed all connections to my mserver.
Problem solved.
________________________________
From: Tapomay Dey
To: Communication channel for MonetDB users
Sent: Tuesday, January 29, 2013 5:56 PM
Subject: Re: caught SIGTERM, starting shutdown sequence
Should I funnel? Plz suggest.
________________________________
From: Tapomay Dey
To: Communication channel for MonetDB users
Sent: Tuesday, January 29, 2013 5:30 PM
Subject: Re: caught SIGTERM, starting shutdown sequence
After around 2-3 minutes the log said
2013-01-29 11:53:10 ERR merovingian[32005]: client error: client XYZ:xyz sent challenge in incomplete block:
2013-01-29 11:53:10 MSG merovingian[32005]: sending process 32010 (database 'msearch_stats_db') the TERM signal
2013-01-29 11:53:11 MSG merovingian[32005]: database 'msearch_stats_db' (32010) has exited with exit status 0
2013-01-29 11:53:11 MSG merovingian[32005]: database 'msearch_stats_db' has shut down
2013-01-29 11:53:11 MSG merovingian[32005]: Merovingian 1.7 stopped
I checked that the monetdbd process did stop successfully. Shouldn't the monetdbd stop command block until the process has ended successfully. I am guessing that my earlier impatient force kills must have caused some kind of corruptions.
Regards,
Tapomay.
________________________________
From: Tapomay Dey
To: Communication channel for Monet DB users
Sent: Tuesday, January 29, 2013 5:23 PM
Subject: caught SIGTERM, starting shutdown sequence
MSG merovingian[32005]: caught SIGTERM, starting shutdown sequence
2013-01-29 11:40:53 MSG control[32005]: control channel closed
2013-01-29 11:41:06 MSG merovingian[32005]: caught SIGTERM, starting shutdown sequence
2013-01-29 11:41:40 MSG merovingian[32005]: caught SIGTERM, starting shutdown sequence
2013-01-29 11:41:51 MSG merovingian[32005]: caught SIGTERM, starting shutdown sequence
Every time I do a monetdbd stop {path_to_farm} I keep getting this in merovingian logs.
But the monetdbd process never ends.
This issue was not there on the same installation of monetdb initially.
This problem started occurring annoyingly frequently after I renamed my last farm and created a fresh one.(with monetdbd not running)
What leads me to stop the farm is that I am unable to connect to the db using mclient(There's another app accessing the DB using JDBC-c3p0). There's is no complaint in the merovingian log. The mclient command just hangs.
I get back access once I force kill the monetdbd using kill -9.
But it gets back into this state after some time.
Running on Feb13 branch some older revision.
Ubuntu 12.10 with lots of resources.
Is this a known issue. Is there a workaround/config or something.
_______________________________________________
users-list mailing list
users-list@monetdb.org
http://mail.monetdb.org/mailman/listinfo/users-list
_______________________________________________
users-list mailing list
users-list@monetdb.org
http://mail.monetdb.org/mailman/listinfo/users-list
_______________________________________________
users-list mailing list
users-list@monetdb.org
http://mail.monetdb.org/mailman/listinfo/users-list
