= Review (NS-99): Bug 1767 - After JMS Broker restart, the bitarchive monitors disapeared =
|| Author || Kåre ||
|| Moderator || Kåre ||
|| State || Closed ||
== Objectives ==
{{{

}}}
== Summary ==
{{{ 
KFC follows up
}}}
'''Total Time Used (Coding,Documentation,Review)''':
{{{ 
Time Used:
KFC: ½md 
HBK: 0.1 md
}}}

'''General comments''':
|| '''Description''' || '''Classification''' || '''Status''' ||


=== Comments on file 'trunk/src/dk/netarkivet/common/distribute/JMSConnection.java', revision 1044 ===
|| '''Lines''' || '''Description''' || '''Classification''' || '''Status''' ||
|| 301 || Don't clear this, it should be reused on reconnect. Or don't use cleanup in reconnect, so we only clear needed state || Minor || OK ||
|| 377 || What should be done on other exceptions? It seems reasonable to retry, but we can't call "onException(...)" since that takes a JMSException as an argument. Should it simply be retried without calling onException? Decision: We always reconnect on non-JMS-exceptions || Minor || OK ||
|| 409 || Here we copy the listeners, to reestablish them after reconnect - but if the reconnect fails, we forget them forever. That seems a little scary. So don't do that! Never clear listeners on reconnect || Minor || OK ||
|| 599 || As above, what should be done on non-JMSExceptions || Minor || OK ||
|| 645 || Again, what about non-JMS-exceptions || Minor || OK ||