[collectd] collectd-ping: will not retry hostname lookup after
octo at verplant.org
Fri Oct 20 09:30:23 CEST 2006
Forwarded message from Simon Richter <sjr at debian.org>
> as collectd starts early, the network is not yet fully configured at
> this point (autonegotiation still running etc.). So name lookup does
> not properly work at this point. The ping plugin then gives up and
> does not record any data.
that problem should be possible to handle: We just try to add the host
in increasing intervals, as Sebastian suggested, too.
> Using the IP address would not solve the problem, as the data is
> recorded in a file that is named after the reverse lookup for the IP
> address, so it would be named differently depending on whether lookup
> is available or not.
Since we need to have a list of hosts within the ping-plugin for the
first point, storing the name that was used in the configfile in there
shouldn't be a big issue. That name could then be used to generate the
filename, preventing problems with changing/RR-reverse lookups as well.
On Wed, Oct 18, 2006 at 02:28:25PM +0200, Sebastian Harl wrote:
> However, this would block collectd. Thus, the call to ping_add_host
> should be moved to the ping plugin's init function (which Imho is the
> correct place anyway) as it is called after forking to the background.
Right, doing the lookup while parsing the configfile is not the best way
of doing this..
I'll have a look at this..
Florian octo Forster
Hacker in training
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Size: 189 bytes
Desc: Digital signature
Url : http://mailman.verplant.org/pipermail/collectd/attachments/20061020/53e184da/attachment.pgp
More information about the collectd