- Products
- Learn
- Local User Groups
- Partners
- More
MVP 2026: Submissions
Are Now Open!
What's New in R82.10?
Watch NowOverlap in Security Validation
Help us to understand your needs better
CheckMates Go:
Maestro Madness
Dear community.
I found the solution sk109236 - "High CPU / process crashes / timeout due to large database / first time operations / load on the machine".
The solution to get better performance is to modify the connection buffer.
"The size of the Communication Buffer is hard-coded to 1MB-4MB depending on the version (For R77.10 and above: 4MB)."
"It is strongly suggested to set FWASYNC_MAXBUF=40000000 (or more, according to what the allocation failed on, i.e. if it failed on 40MB, then set to 80MB)"
Regards
Roman
Increasing a buffer (and yes, there are a lot of buffers in CP GWs and SMS !) adds to memory consumption, so when buffers are very large, swapping will occur and slow down the unit...
I believe FWASYNC_MAXBUF is used by the fwm process on a SMS to buffer incoming operations that fwm needs to perform. In R77.30 management and earlier, fwm had far too many responsibilities and was mostly single-threaded. This was a bad combination and led to fwm daemon crashes under heavy load which would eject the administrator from the SmartDashboard and all other GUI tools.
fwm still exists in R80+ management, but mainly to service the older legacy GUIs like the Smartview Monitor and SmartDashboard and a few other operations. Given the wholesale changes in R80+ management, fwm has much less to do these days and has been pretty stable from what I've seen; the new cpm daemon has taken over most of fwm's prior responsibilities. As such I wouldn't worry too much about increasing this buffer for fwm in R80+ management, unless you are actively seeing warnings about running out in $FWDIR/log/fwm.elg.
--
Second Edition of my "Max Power" Firewall Book
Now Available at http://www.maxpowerfirewalls.com
Thank you for the answers.
We are using R77.30 Take 292.
$FWDIR/log/fwm.elg is not there.
I increased the buffer to 80 MB. It seems to work better, now.
Regards
Roman
Next thing we did was to delete the caching file $FWDIR/conf/CPMILinksMgr.db
It was 60 Gb and maked more free RAM space.
128/125 Gb used to 128/70 Gb
Hi Roman,
we are using these parameters since more than 1 year on R77.30 with HFA T225 (and earlier). We configured it on the MDS and the GWs and the last time I read this SK the limit was about 64MB and this is what we configured.
To be honest I don't know what we want to solve in the past with this buffer. I did not notice any bad side effects.
Regards
Leaderboard
Epsum factorial non deposit quid pro quo hic escorol.
| User | Count |
|---|---|
| 19 | |
| 16 | |
| 7 | |
| 5 | |
| 5 | |
| 5 | |
| 4 | |
| 4 | |
| 4 | |
| 4 |
Tue 16 Dec 2025 @ 05:00 PM (CET)
Under the Hood: CloudGuard Network Security for Oracle Cloud - Config and Autoscaling!Thu 18 Dec 2025 @ 10:00 AM (CET)
Cloud Architect Series - Building a Hybrid Mesh Security Strategy across cloudsTue 16 Dec 2025 @ 05:00 PM (CET)
Under the Hood: CloudGuard Network Security for Oracle Cloud - Config and Autoscaling!Thu 18 Dec 2025 @ 10:00 AM (CET)
Cloud Architect Series - Building a Hybrid Mesh Security Strategy across cloudsAbout CheckMates
Learn Check Point
Advanced Learning
YOU DESERVE THE BEST SECURITY