[
https://issues.apache.org/jira/browse/CASSANDRA-959?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jonathan Ellis resolved CASSANDRA-959.
--------------------------------------
Resolution: Not A Problem
Assignee: (was: Pavel Yaskevich)
It's important for the data nodes to know how long the coordinator will wait
for a reply so we can drop hopeless requests under overload conditions. I
don't see a good way to do that, and also implement variable rpctimeouts.
(Where "good" includes "not breaking RPC compatibility, e.g. by explicitly
including the timeout in the message.")
Timeouts indicate an overload scenario, so tuning it to time out earlier for
some messages is nice but not critical (because the real fix is, make your
workload less bursty and/or add more capacity).
> Allow different timeouts for different classes of operation
> -----------------------------------------------------------
>
> Key: CASSANDRA-959
> URL: https://issues.apache.org/jira/browse/CASSANDRA-959
> Project: Cassandra
> Issue Type: New Feature
> Components: Core
> Reporter: Jonathan Ellis
> Priority: Minor
> Fix For: 0.8
>
>
> Currently we have one rpc timeout for intra-node operations. But applying
> the same timeout to "read one row", "read multiple rows [by key]," and "range
> query multiple rows" feels like an increasingly uncomfortable fit. (See e.g.
> CASSANDRA-919.)
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
