Hi,
I'm seeing odd results from 112's and 153's on my network. OSPF is used on the network and as the routes come and go the free memory of the nodes going down till it's released again. The problem I'm seeing is the memory goes so low on the 112's and 153's before being released that you can't launch a terminal. Other erratic behaviour has also occurred at this point like losing its OSPF neighbours.
thanks,
Mick

We are graphing memory and CPU usage of these boards with MRTG currently and Mehall has set up a second RB153 with 2.9.29.

The behavior with the console crashing due to lack of memory is something that exists as far back as 2.9.24, but 2.9.31 seems to be specificly bad as memory usage is constantly growing to a system standstill.

/Marlow

Hi,

the graphs can be found at http://valhalla.sca.airwire.ie/traffic/

If you check the graphs you can see how the pattern differs.

RB112/RB153 at 2.9.31 load the memory and crash
RB153 at 2.9.29 loads the memory up, but never comes to a standstill
RB532 at 2.9.31 loads the memory up, but only very slight, doesn't crash
a WRAP (ngeode platform) at 2.9.31 doesn't seem to have that behavior at all.

/Martin

my graphs for the 532's look very similar...

http://forum.mikrotik.com/viewtopic.php?t=11437