I'm running CAS 3.5.1 against an Oracle10 database. There are about 34,000 
tickets in the ticketgrantingticket table. When the JPA ticket cleaner runs, 
after about 90 seconds I see the following in the logs

13:03:32.687 [scheduler_Worker-3] ERROR o.h.util.JDBCExceptionReporter - 
ORA-01555: snapshot too old: rollback segment number  with name "" too 
small...caused by javax.persistence.QueryTimeoutException. Invocation of method 
'clean' on target class [class 
org.jasig.cas.ticket.registry.support.DefaultTicketRegistryCleaner] failed; 
nested exception is javax.persistence.QueryTimeoutException: could not execute 
query

About a month ago I resolved these same errors by removing an abandoned 
connection timeout (which was set to 60s) on the datasource, which allowed the 
cleaner to complete in slightly over a minute. 

However, I've no idea where this 90s timeout might be configured, assuming 
that's even the source of the problem. I've also had to increase the available 
heap space on the server because all these tickets are being loaded into 
memory. Nevertheless, CAS still seems to be going down every day or so because 
this cleanup job is failing. Does anyone have any ideas on what to do? 

Thanks much,
Tim
-- 
You are currently subscribed to [email protected] as: 
[email protected]
To unsubscribe, change settings or access archives, see 
http://www.ja-sig.org/wiki/display/JSG/cas-user

Reply via email to