Write contention workaround in 1.1 forbroken aggregate_status

[Keith Weinberg]

I've found a work-around to the write-contention problems that I was having
with 1.1

I've disabled aggregate_status  (in nagios.conf, setting it to 0).

It seems that there is a bug in this function, that causes write contention
and significantly slows down the system (and even may cause deadlock) -
Something about the aggregator seems to be broken (I haven't gotten to
peruse the code yet, but there is certainly some issue)

In order to keep my filesystem (ext3) humming along fast enough to keep up
with the non-aggregated updates, I now mount it with "noatime" so that there
is less work for the filesystem to do.

Now my "nagios-owned" process queue ( ps auxww | grep nagios | wc -l) hovers
between 4-20 as opposed to being in the 200-500 range and my nagios-schedule
is only minutes behind (as opposed to hours).

Anyone else have similar experiences?

/<
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://www.monitoring-lists.org/archive/users/attachments/20030929/7c82af36/attachment.html>