On 4/27/09 2:55 PM, "Ariel Rabkin" <[email protected]> wrote:
> A couple notes on the current admin guide.
>
> 1) There's a reference to an internal Yahoo! hudson under "Chukwa
> Binary". Probably should be excised.
This should be renamed.
> 2) Do I need to download tomcat and stick it in /opt? If not, how
> does HICC get ahold of it?
Yes, download tomcat and extract to opt, and you should copy hicc.war to
apache-tomcat/webapps, then copy descriptors, views from
chukwa/opt/apache-tomcat/webapps/hicc to
Chukwa/var directory.
I will ask Corinne to include this in the documentation.
> 3) What happens if I add multiple databases to jdbc.conf ?
jdbc.conf supports multiple databases with this syntax:
<cluster>=jdbc:mysql://<host>:<port>/database?user=<user>&password=<password
>
If the data processor node contains multi-line jdbc.onf, it will run
aggregation and down samples on the defined databases.
> 4) Do we need to give the whole my.cnf in the admin guide? It breaks
> up the flow of the document. Can't we just say, "look in
> CHUKWA/opt/..."?
+1 on this improvement. CHUKWA-191 filed.
> 5) We still have bin/start-agents.sh and start-collectors.sh Do these
> still work? Are they being maintained? Should they be documented?
Both of those should work, but we are not using them in Yahoo nor testing
them.
Regards,
Eric
