Cisco found below bug and advised to upgrade.
1. CSCut24730 - MSE 10.X - ERROR com.cisco.mse.persistence.jdbc.JdbcQueryResultSet
This bug causes the system to overutilize CPU and some records might be dropped, as the system has no resources to process them properly.
We can see the matching log messages exactly at the time you observed the issue (significant change in user statistics):
2016-05-14T01:00:02,212 [analytics:aggregation:AggregationExecutor:105] WARN com.cisco.mse.persistence.UserTransaction - Not committing transaction since usage count (2) != 1 2016-05-14T03:08:35,275 [Thread-36] ERROR com.cisco.mse.persistence.jdbc.JdbcQueryResultSet - Could not retrieve binary data. (Not a postgres db?) : ERROR: large object 831420821 does not exist 2016-05-14T03:31:35,496 [Thread-35] ERROR com.cisco.mse.persistence.jdbc.JdbcQueryResultSet - Could not retrieve binary data. (Not a postgres db?) : ERROR: large object 831463875 does not exist
2. CSCuu06749![]()
10.2: Path analysis report generation fails due to 400 bad request
It’s a bug in path analysis, that causes the system to calculate client path in a wrong way. It can affect the client statistics (where exactly the client was seen, so where should the system count it in).
This is the marker for this bug:
ERROR com.cisco.mse.analytics.aggregation.collection.AggregationRunner - AR-008: Received new notification with one area, no parents
Both bugs are fixed in 10.2.0 version. Both bugs were internal, so I’m not sure if you can see them right now – they may become visible a bit later on.
I advise you to upgrade to 10.2.1 to be sure your system is free of these bugs and will work reliably in the future.
