On Thu, Mar 13, 2008 at 10:43:20AM -0500, Joe Wells wrote:
> This seems backwards to me, but no big deal. I thought that a counter
> would be used for straight-ahead reporting of the number of widgets
It probably is backwards. Remember this comes from the folk who designed
SNMP. Anyone that thinks
On Thu, 2008-03-13 at 11:41 +1100, Craig Small wrote:
> On Mon, Mar 10, 2008 at 05:08:06PM -0500, Joe Wells wrote:
> > Yup, it's a counter. I'm re-doing it as gauge to see what happens. Looks
> > better already, though.
> >
> > I thought that this would be a counter, and the JFFNMS docs list it a
Hi
I'm not sure if that blockdev command helped or not...it didn't make things
worse, that is for sure. One thing that DID help quite a bit, was to edit
the Physical Interface (Interface Type) and reduce the Pings to Send from 50
to 0 which REALLY helped bring the load down. Thanks again. Lookin
That looks right, as Craig guessed it is the pipe to the rrd file. It
definitely looks like you are having problems with your rrd files.
How did the "blockdev --setra 64" go, did it help at all?
Cheers
Martin
... wrote:
> Here is my output from that command;
>
> [EMAIL PROTECTED]:/opt/jffnm
Here is my output from that command;
[EMAIL PROTECTED]:/opt/jffnms/engine# ls -l /proc/2854/fd/7
lr-x-- 1 jffnms jffnms 64 2008-03-13 09:42 /proc/2854/fd/7 ->
pipe:[11576087]
[EMAIL PROTECTED]:/opt/jffnms/engine# lsof | grep 11576087
php2854 jffnms7r FIFO0,6
