Slack digest for #general - 2018-11-23

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

2018-11-22 13:43:08 UTC - Christophe Bornet: Hi. Wondering if there are any recommendations about the network quality for BookKeeper and Pulsar. Especially, is there any impact when Bookies are spread over multi-datacenter with network latency ? Is there an effect on throughput ? Stability ? Only write/read latency ? In that case, what would be the max recommended network latency between data centers ?
----
2018-11-22 19:02:45 UTC - Yuvaraj Loganathan: Is it possible to increase the number of partitions from 4 to 8 ? for a topic  How does the client will behave. Will the client start new thread to consume from the new partitions ?
----
2018-11-22 19:08:26 UTC - Sijie Guo: BookKeeper has been deployed over multi datacenters. The only impact will be latency. If your application is asynchronous, your throughput will not be impacted. I have seen bookkeeper deployed in production in a 100~200 network latency enivronment. but it is probably okay to deployed to a environment has higher latency.

but you might need to tune settings such as connect timeout, write/read timeout. you should also enable region aware placement policy to ensure data is stored in multiple datacenters.
----
2018-11-22 19:09:53 UTC - Sijie Guo: yes. you can update the partitions via : <http://pulsar.apache.org/docs/en/admin-api-partitioned-topics/#update>

but I think currently you need to restart client to refresh the metadata.
----
2018-11-22 19:58:26 UTC - Christophe Bornet: :+1:. thanks for the feedback
----
2018-11-23 03:13:48 UTC - anhldbk: @anhldbk has joined the channel
----