Spectrum OneClick showing incorrect number of alarms. The alarm counts keep increasing until a restart of tomcat one the OneClick system. However, after the restart, they continue to grow.
Release : Any
Component : Spectrum OneClick
Possible network issue.
Check the contents of the $SPECROOT/tomcat/logs/catalina.out (Linux) or stdout.log (Windows) for the following error:
Nov 20, 2020 16:14:39.475 (OrigAlarmEvents) (ALARMS) - BackEndAlarmDataModel getting events for 800 alarms from SpectroSERVER1
Nov 20, 2020 16:18:14.043 (Thread-40) (MDL) - techm: Callback notifications were missed from landscape SpectroSERVER1 - resynching cached alarms.
Run a sniffer trace on the OneClick system and check for packets with "TCP previous segment not captured" errors.
"TCP previous segment not captured" means that a gap was detected, either because the packet was lost on the way from sender to receiver, or that the capture device recording the packets wasn't fast enough to deal with the incoming flood of packets."
"If packet N is marked with previous segment not captured, it means that in the capture there is no packet from the same TCP session whose seq + length would match the seq of packet N. The most typical reason is packet loss and/or late start of capture, which is the reason why the wording in question is used."
Run "netstat -s" on the OneClick system and look for error similar to the following:
797564 segments retransmited
131208 bad segments received.
Check the status of the network between the OneClick and SpectroSERVER systems.