Thanks for the update. On Tue, Jul 25, 2017 at 10:13 PM Paulo Motta <[email protected]> wrote:
> This is address on 3.11.1 on CASSANDRA-13641. Workaround for now is > probably truncating system.prepared_statements before restart of node > as being done for now. > > 2017-07-25 11:12 GMT-05:00 Taylor Cressy <[email protected]>: > > +1 bump. > > > > We are experiencing the same issue > > > > On Jul 25, 2017, at 08:02, Anshul Rathore <[email protected]> > wrote: > > > > Hi > > > > We recently upgraded to cassandra 3.10 , one of issue which we are > facing is > > cassandra node restarts are taking very long times. > > > > On checking the debug logs, it appears cassandra saves all prepared > > statements ever executed in system.prepared_statements table, and on > startup > > tries to load all the prepared statement on startup and somtimes takes > hours > > to restart or even worse goes in gc cycle and never restarts > > > > I can see there are millions of prepared statements , so currently we > > trucating system.prepared_statements before restart of node. > > > > Also the large number of prepared statemets are due to some tables have > over > > 100 columns , so number of permuations of query statements are huge. > > > > Anybody has any idea on this , how to avoid or fix this ? > > > > Thanks, > > Anshul > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] > >
