On Oct 10, 2007, at 9:19 PM, Jeff Terrell wrote: > I am going to be setting up a couple more slaves soon, so hopefully > I'll be able to pay you back with some documentation improvements > soon. :-)
I looked over the "smokeping_master_slave" document on the web page, and there are several important things missing. I think we could add a new section, "troubleshooting", to cover these points. Here is what I have so far. Comments and corrections are welcome. Troubleshooting o The host names are expected to be short names. Change fully- qualified slave names in the configuration to short names. Do this even if the master and slave are in different domains. (The master looks up configuration based on the name the slave uses to identify itself. As long as there are not two slaves with the same short name, everything should be fine.) You might also need to change the hostname of the slave so that it reports a short name with the `hostname` command. (Or, install Smokeping version X.X.X or later and use the --slave-name option, e.g. '--slave-name=`hostname -s`'.) o The secrets file on the master is of the form "slave1:secret1 \nslave2:secret2", but the secrets file on the slave should have only the secret of that slave. For example, the secrets file on slave1 will contain only "secret1". o Make sure that smokeping.cgi runs on the master with permissions that allow it to write to the data directory. o Beware selinux, which could cause strange and unreported permissions problems. Notes: - The hostname patch needs to be in the next version. Then, update the "X.X.X" to reflect the new version number. - Speaking of the hostname patch, I think that might need to be tested. I might be misremembering this, but I think I ended up using the patch on the master, but not on the slaves. Or something like that... Stephen, do you have anything to add to this list? Thanks, -Jeff T. _______________________________________________ smokeping-users mailing list [email protected] https://lists.oetiker.ch/cgi-bin/listinfo/smokeping-users
