I've never seen a JVM crash that was polite enough to run shutdown hooks first, but it's worth a try.
On Wed, Oct 12, 2011 at 3:27 PM, Erik Forkalsrud <eforkals...@cj.com> wrote: > > My suggestion would be to put a recent Sun JVM on the problematic node and > see if that eliminates the crashes. > > The Sun JVM appears the be the mainstream choice when running Cassandra, so > that's a more well tested configuration. You can search the list archives > for OpenJDK related bugs to see that they do exist. For example: > http://mail-archives.apache.org/mod_mbox/cassandra-user/201107.mbox/%3CCAB-=z42ihihr8svhdrbvpfyjjysspckes1zoxkvcje9axkd...@mail.gmail.com%3E > > > - Erik - > > > On 10/12/2011 12:41 PM, Ashley Martens wrote: > > I guess it could be an option but I can't puppet the Oracle JDK install so I > would rather not. > > On Wed, Oct 12, 2011 at 12:35 PM, Erik Forkalsrud <eforkals...@cj.com> > wrote: >> >> On 10/12/2011 11:33 AM, Ashley Martens wrote: >> >> java version "1.6.0_20" >> OpenJDK Runtime Environment (IcedTea6 1.9.9) (6b20-1.9.9-0ubuntu1~10.10.2) >> OpenJDK 64-Bit Server VM (build 19.0-b09, mixed mode) >> >> This may have been mentioned before, but is it an option to use the >> Sun/Oracle JDK? >> >> >> - Erik - >> > > > > -- > gpg --keyserver pgpkeys.mit.edu --recv-keys 0x23e861255b0d6abb > Key fingerprint = 0E9E 0E22 3957 BB04 DD72 B093 23E8 6125 5B0D 6ABB > > -- Jonathan Ellis Project Chair, Apache Cassandra co-founder of DataStax, the source for professional Cassandra support http://www.datastax.com
