Re: pmclusterd versus other solutions

[fche@xxxxxxxxxx (Frank Ch. Eigler)]

jhanson wrote:

> [...]  And regardless of the issues here I was interested in what
> other people do with pcp to monitor cluster nodes.  [...]

That's a good question.  PCP's one-host/archive-at-a-time mentality
percolates, with only a few tools allowing limited co-mingling of data
for a small number of machines.  Our archive files, while well tuned
for efficient initial logging from an individual host, are an obstacle
to joint or ad-hoc analysis across time spans or hosts.  The
multi-archive work is a step forward, but there is a long way to go.

- FChE