[Spread-users] Spread crashing randomly & logging impact on performance

[Matt Garman]

We're using spread as the "middleware" for some custom
systems-monitoring tools.  These tools run 24x7.

The load is fairly high, but not outrageous: an upper bound on the
size and volume of messages is about 1000 100-byte messages per second
(but with huge variance).

Our setup has two segments: one with two linux servers, and another
with one linux server and four linux virtual machines.

What we're finding is that our spread deamon is crashing every 10 hours or so.

I have spread set up to restart if it crashes.  Unfortunately, I made
a configuration mistake and had my logfiles overwritten with each
re-start (was using stdout/stderr re-direction to a file).  So my logs
don't have any useful information.

At any rate, I have two questions: (1) has anyone seen similar
stability issues as described above, and (2) besides fixing my logging
problem, where is a good place to start looking for the cause of this?
 Also, what are the performance implications of turning up the logging
options?

Thank you,
Matt