Slack digest for #general - 2018-07-28

[Apache Pulsar Slack <ap...@gmail.com>]

2018-07-27 14:05:26 UTC - Peter CW: @Peter CW has joined the channel
----
2018-07-27 17:10:46 UTC - Poule: Let's say I have a working cluster for months where Bookies use EBS to store data. EBS crashes like in 2012 and I have to restore my bookies from a S3 copy from 24 hours ago. Let's say my Zookeeper nodes did not crash and so are up-to-date, (data in Zookeeper is current) but now my data in Bookies is missing the last 24 hours. Will the discrepancy in the data in ZK and my Bookies cause me problems? Will Pulsar still work in such a scenario? If not, what do I have to fix?
----
2018-07-27 17:27:07 UTC - Sijie Guo: Your consumers might be stuck in reading since there will be missing segments. However there are admin tools to move cursors and skip those missing segments. So in general you are fine. 

Also There is a feature called tiered storage is coming in 2.1 release, you might be interested in. Pulsar will manage moving segments to cloud storage like s3 if you enable that feature. 
----
2018-07-27 17:29:16 UTC - Daniel Ferreira Jorge: When I'm consuming from a partitioned topic, how can I, inside the message listener, get the partition number of the message that is being consumed came from?
----
2018-07-27 17:42:34 UTC - Poule: @Sijie Guo ok thanks
----
2018-07-27 17:43:44 UTC - Sijie Guo: @Daniel Ferreira Jorge right now, I think you can cast MessageId to PartitionedMessageIdImpl to get the partition number.
----