Archive Manager is disconnecting or shutting down on its own
Article ID: 253071
Updated On:
Products
Issue/Introduction
Sep 21 10:49:21 : ArchMgr has successfully advertised CORBA Event Service.
Sep 21 10:49:21 : ArchMgr is now ready on port 0xbafe, precedence 10Sep 21 10:55:49 : ArchMgr has received shut down signal - scheduling shut down
Sep 21 10:55:49 : Closing all client connections...
Sep 21 10:55:49 : Stopping CORBA communication...
Sep 21 10:55:58 : Waiting for 217281 pending database writes to complete................................................................................................................................................................................................................
MySQL is going down.
Alternatively you may see an alarm on the VNM noting The Internal event log queue is overflowing.
Environment
Release : 21.2
Cause
Trap and event rate is too high.
Resolution
This is ultimately a Events volume issue. Try and reduce the incoming trap rate.
If that cannot be done in my-spectrum.cnf file add:
net_read_timeout=60
connect_timeout=30
Save and restart mysql service.
Also add more CPUs and Memory if possible.
Reduce overall Event volume where possible.
You can increase the internal queue size by editing $SPECROOT/SS/.vnmrc and add:
max_internal_log_queue_size=20000
Cycle the SS.
If the issue still exists you can double the mysql values set. So you can edit my-spectrum.cnf and set:
max_allowed_packet=32M
net_read_timeout=120
connect_timeout=60
Cycle processd to reset mysql.
If that still does not resolve the issue you must reduce the trap and event rate.
Additional Information
If using the above settings does not resolve this, you can try adding the following entry to the $SPECROOT/SS/.vnmrc file:
max_internal_log_queue_size=20000
In additions, double the mysql values set. So you can edit my-spectrum.cnf and set:
max_allowed_packet=32M
net_read_timeout=120
connect_timeout=60
Stop the SpectroSERVER and processd.
Start processd and then the SpectroSERVER
Feedback
