Mxosrvrs are restarted immediately by DCSServer as soon as it exits to ensure the configured count of mxosrvr are always available for connection. So, just checking for count from dcscheck command wouldn't be sufficient to conclude that the mxosrvrs didn't exit in the midst of the query.
Selva -----Original Message----- From: Eric Owhadi [mailto:[email protected]] Sent: Friday, September 22, 2017 11:21 AM To: [email protected] Subject: RE: T4 driver error. What can this be Exception:org.trafodion.jdbc.t4.TrafT4Exception: Server aborted abnormally or Connection timed out I did not check for core files because all MXOServers were still up and running after I received more than 60 of these timeout. I would expect 60 MXOSever down if root cause is CORE related. Each query is using a serial plan, so only run on MXO. No esp involved, Eric -----Original Message----- From: Selva Govindarajan [mailto:[email protected]] Sent: Friday, September 22, 2017 1:18 PM To: [email protected] Subject: RE: T4 driver error. What can this be Exception:org.trafodion.jdbc.t4.TrafT4Exception: Server aborted abnormally or Connection timed out Did you check if there are any core files dumped on the server side? If there were any core files, please check the ulimit for core file size. ulimit -a If you know the pid of the server process, you can also check if the process ended in the monitor map file at $TRAF_HOME/tmp directory. Selva -----Original Message----- From: Eric Owhadi [mailto:[email protected]] Sent: Friday, September 22, 2017 10:02 AM To: [email protected] Subject: T4 driver error. What can this be Exception:org.trafodion.jdbc.t4.TrafT4Exception: Server aborted abnormally or Connection timed out Hi Trafodioneers, I am playing with a spj that uses T4 connection to perform long running upsert operation from one table to another. It worked fine on small tables, but testing it on very large table, where each upsert takes a lot more time, I am getting these: "Thread 62, partition 2451039, Exception:org.trafodion.jdbc.t4.TrafT4Exception: Server aborted abnormally or Connection timed out". I suspected something along socket timeouts not expecting work to be so long on the DB side, and thought I'll just have to configure some jdbc default time out and be OK with it. But when I look at the documentation, the following properties have default that looks OK, (not sure about the connectionTimeout?): connectionTimeout -> -1 (Use the ConnTimeout value set on the server data source.) maxIdleTime 0 -> means specify no limit networkTimeout 0 -> means no network timeout specified So what could this be? Are these documented defaults not accurate? Thanks in advance for the help, Eric
