[
https://issues.apache.org/jira/browse/CASSANDRA-7222?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14000093#comment-14000093
]
Tyler Hobbs commented on CASSANDRA-7222:
----------------------------------------
[~mishail] we _can_ rely on duration being set one the request is complete
(from the coordinator's perspective). In the case of timeouts, it's true that
events can be added after {{duration}} is already set, but cqlsh has no way of
knowing it should wait for those (or how long, etc), and they're not terribly
important. Retrying the fetch until {{duration}} is set should be sufficient.
> cqlsh does not wait for tracing to complete before printing
> -----------------------------------------------------------
>
> Key: CASSANDRA-7222
> URL: https://issues.apache.org/jira/browse/CASSANDRA-7222
> Project: Cassandra
> Issue Type: Bug
> Components: Tools
> Reporter: Tyler Hobbs
> Assignee: Mikhail Stepura
> Fix For: 2.0.9
>
>
> cqlsh currently sleeps 0.5 seconds before fetching trace info. Instead, it
> should fetch the {{system_traces.sessions}} row in a loop until the
> {{duration}} column is set (maybe with a total timeout of 30s). After the
> {{duration}} column is set, it's safe to fetch the rows from
> {{system_traces.events}}.
> This is already fixed in 2.1 by the switch to the python driver, we just need
> to fix 2.0.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
