R81.20 - How to increase the log rate from management

[gperes]

I have a virtual appliance with 64GB of RAM and 24 CPUs the load is about 0.34, but the log receive rate is 300 and my peak is about 2200. I need to increase the log receive rate but i don't want to mess the configuration of the appliance. Which parameter can I change to increase this setting?

ERROR : In 'VM' machine (RAM:63 CPU: 24), the average Logging rate is 2111 logs per second, while supported logging rate per machine size is 300 logs per second.

[Timothy_Hall]

The is no hard log rate limit that you can increase.  Assuming that the single-threaded fwd daemon which handles the log reception is not hitting 100% on a single CPU (unlikely at those rates), the next bottleneck is usually disk write speed.  Virtual Machines are notorious for having oversubscribed disk paths being shared by other I/O-hungry VMs such as database servers.  You could do a simple command like this from expert mode to benchmark your disk write speed:

sync; dd if=/dev/zero of=/var/log/whatever bs=1M count=1024; sync

Beyond that check out sk104848: Best Practices - Performance Optimization of Security Management Server installed on VMwar...

[gperes]

I already did the optimizations from the SK, but didn't increase the rate. Till now the only response from the TAC is to increase the memory to 128GB. I can't justify the increase to 128GB of RAM, if I'm not using half of the memory allocated.

[the_rock]

Can you send the output of free -h command? Also cpview where it shows memory allocation.

Andy

[gperes]

free -h
total used free shared buff/cache available
Mem: 62G 10G 1.7G 966M 50G 50G
Swap: 63G 228M 62G

[gperes]

the disk test give me a rate of 1.3 to 1.5GBps writing on the disk on /var/log

[gperes]

I saw the I/O, but it's in 1ms the statistics, till now I could't resolve the issue.

[the_rock]

Im fairly sure Tim's response is right, but you can verify with TAC if you need to get an official answer, just to be sure.

Best,

Andy