server.shutdown not working
Hi,
Although I have stopped the server by ./server.shutdown, ps -ax says Jazz server is still running; next ./server.shutdown leads to an exception (the connection refused) and it is not possible to connect to the server. I'm afraid if I kill the process the data will be lost. What is wrong? Jazz1.0Beta2, Linux redhat Roman |
5 answers
Hm, actually the process disappeared after 5-10min...
Regards, Roman "Roman Smirak" <roman.smirak@tietoenator.com> wrote in message news:fnsf7a$ci1$1@localhost.localdomain... Hi, |
Hi Roman,
It is good news that the server eventually stopped. That said, we want to ensure that shutdown is responsive, and there have been prior reports about the server failing to shutdown completely. If you find that you have to kill the process, the only potential for data lost would be a user or background transaction that has not yet completed. The system should be able to recover fine on restart. If this happens again, we can help diagnose the problem if you can capture a Java trace and attach it to a new work item, assigned to the Repository component. To capture the trace, on Windows type CTRL+BREAK at the console window; on Linux run "kill -3 <pid>". -- Ritchie ********************************************** Ritchie Schacher Jazz Repository component lead IBM/Rational RTP Lab schacher@us.ibm.com "Roman Smirak" <roman.smirak@tietoenator.com> wrote in message news:fnsksj$f13$1@localhost.localdomain... Hm, actually the process disappeared after 5-10min... |
Thanks Ritchie, when it comes to Derby DB, do you see any potential risk
there? Recently I had a discussion with a colleague of mine who pointed to a Derby datafile corruption they experienced even in case of regular shutdown. Have you heard about/experienced it? Regards, Roman "Ritchie Schacher" <schacher@us.ibm.com> wrote in message news:fnu7kb$pbs$1@localhost.localdomain... Hi Roman, |
Roman,
A quick search of our bug database turned up the following reports: Team Concert crash make the database(derby) inconsistant (29378) (web) Invalid checksum on Page Page(0,Container(0, 11057)), expected=135?296?395, on-disk version=644?144?805 (40866) (web) The first turned out to be a different problem, and in the second case, we are not sure if an ungraceful shutdown was the cause. Generally we'd hope that derby would be crash tolerant, but as with a commercial db, it's wise to take nightly backups. I can tell you that before the Jazz team scaled up to db2 we used to host our own development on derby and we never experienced data corruption, even though we had our share of server crashes :-(. We also stopped the server nightly for backups. -- Ritchie ********************************************** Ritchie Schacher Jazz Repository component lead IBM/Rational RTP Lab schacher@us.ibm.com "Roman Smirak" <roman.smirak@tietoenator.com> wrote in message news:fnv0m0$4be$1@localhost.localdomain... Thanks Ritchie, when it comes to Derby DB, do you see any potential risk |
Ok, so to get the backup you stopped the server and copied db files and then
started server again - is this correct? Regards, Roman "Ritchie Schacher" <schacher@us.ibm.com> wrote in message news:fnvpmo$f3s$1@localhost.localdomain... Roman, |
Your answer
Dashboards and work items are no longer publicly available, so some links may be invalid. We now provide similar information through other means. Learn more here.