<URL: http://www.dcache.org/rt/Ticket/Display.html?id=2929 >
Hi,
Last night a pool (1.8.0-12p6 on Solaris10) got a bad shape after the traditionnal "sleepycat.util.RuntimeExceptionWrapper" error:
05/06 21:14:51 Cell(pool-cms-hpssdata-sn020a@ccdcatsn020Domain) : Flush thread : loop interrupted :
com.sleepycat.util.RuntimeExceptionWrapper: (JE 3.2.44) Channel closed, may be due to thread interrupt
05/06 21:15:51 Cell(pool-cms-hpssdata-sn020a@ccdcatsn020Domain) : Flush thread : loop interrupted : java.lang.IllegalArgumentException: Is
already active
The last line was repeated thousands of times until the FS got full. Is this error related to the previous one? I had never seen it before.
Thanks
Lionel
Oh, it seems flush jobs receive thread interrupts too... I didn't
protect those from killing the Berkeley DB, so that's something I now
have on my todo list.
It is likely that the IllegalArgumentException is triggered as a result,
but I will need to look through the flush code to make sure.
Cheers,
