| time |
nick |
message |
13:48 |
<chirino_m> |
jstrachan: wish the yahoo guys joined #zookeeper |
13:51 |
<jstrachan> |
which irc server? |
13:51 |
<jstrachan> |
is there a zookeeper room? |
13:52 |
<jstrachan> |
have nearly got the WriteLock working nicely with server connection failures - its amazingly hard! :) - nearly there though |
13:52 |
<chirino_m> |
jstrachan: we could create one but it would just be and you :) |
13:53 |
<chirino_m> |
jstrachan: yay |
13:54 |
<jstrachan> |
:) |
13:54 |
<jstrachan> |
yeah |
13:58 |
<chirino_m> |
once your done there you could work on the ReadWrite Lock :) |
13:58 |
<jstrachan> |
yeah :) |
13:59 |
<jstrachan> |
I've hacked it so that most of the code should just work for ReadWrite lock |
13:59 |
<jstrachan> |
e.g. ZNodeName is sortable by read/write locks first, then by sequence number etc |
14:02 |
<chirino_m> |
just commented on the jira.. is it possible to make that guy implement j.u.concurrent.locks.Lock |
14:04 |
<chirino_m> |
jstrachan: btw I don't think that compare will work for rw lock. |
14:05 |
<jstrachan> |
you'd wanna see all the current locks sorted in order no? |
14:06 |
<chirino_m> |
yeah but I don't want the read/write locks to jump ahead |
14:06 |
<chirino_m> |
and that's what would happen in the compare |
14:06 |
<chirino_m> |
since you compare the prefix first. |
14:06 |
<jstrachan> |
I mean - you could find all the write locks, or all the read locks etc |
14:07 |
<jstrachan> |
or we could put all the read locks in a different collection to the write locks if needed :) |
14:07 |
<chirino_m> |
no |
14:07 |
<chirino_m> |
you want them in the same. |
14:07 |
<chirino_m> |
and have the same sequence |
14:07 |
<jstrachan> |
but figured we could search the sorted list to find the writes or read locks |
14:07 |
<jstrachan> |
ah you want 'em sorted in sequence order, not prefix order? |
14:07 |
<chirino_m> |
ya |
14:07 |
<jstrachan> |
ah ok! |
14:08 |
<jstrachan> |
that'd be easy to do too |
14:08 |
<chirino_m> |
so if there are reads ahead of a write they all can get the lock. |
14:08 |
<chirino_m> |
but once a write is ahead of the reads only the 1st write lock gets the lock. |
14:09 |
<jstrachan> |
yeah |
22:48 |
<Drizzt321> |
is it possible for different consumers to consume the same Message from a Queue? |
22:48 |
<Drizzt321> |
Or for the same consumer to get the same message more than once from the same queue? |
22:49 |
<Drizzt321> |
because I'm seeing a queue that goes has the number of messages pending as less than 0 |
22:50 |
<Drizzt321> |
and the messages received is greater than the messages sent |
22:50 |
<Drizzt321> |
by the same exact amount |