Open Support-DataExpertise opened 3 years ago
Hello
Just now i see no any reason to move production server to UniFi Controller v6.
And i have no free equipment for test installation building.
May be CPU utilization growth is happened because some utility misconfiguration exist? Small "cachemaxage" value for example.
I have installed v6 on my home notebook, and maked query flood by UniFi Miner with some simple queries (site discovery, get info) to "empty" controller. And, didn't see no freezes or slow replies.
Also, i copied unifi_miner.pl to just_print_something.pl and inserted print & exit commands right after Perl libraries initialization. This "empty" script on flood scenario maked the same CPU load that original.
I guess that affected system analyze only can answer the question "whats wrong".
Hi,
Thanks for your reply, it's really strange because, the day I've activated the script, the CPU utilization bumped up quite significantly :
And here is the load on cpu on Zabbix-Proxy (trigger at 1.5 by default)
Here is also the result of top command on my unifi server
I have no idea.
Debug or profiling can make this mystery more clear.
Hi,
The version 6 of unifi controller is now available, we ave updated some of our controller to it but I don't if it's that but those VMs have a huge CPU utilization when executing the script. We still have metrics that works BTW so the API doesn't seems to have change a lot. When do you think you will update for support of version 6 ?
Thanks for the reply and great work !