OPNsense Forum
Archive => 17.7 Legacy Series => Topic started by: BertM on August 21, 2017, 01:33:53 pm
-
We have a number of OPNsense instances running that, until recently, were running OPNsense 16.7.14_2.
They had not been updated to 17.1 because of some issues we had with Mobile IPSec connections.
After testing showed that these issues were solved in 17.7, I updated the first 5 of them to the latest version.
Obviously, SNMP has now changed to a plugin, so I installed the os-snmp plugin.
On all these OPNsense installations I noticed that the processor utilisation flat-lined to 100%, although performance did not seem to be influenced much.
This high processor utilisation went away in all cases when I unchecked the box for "Host Resources" on the Services:SNMP page.
Note that:
- The issue exists on all 5 OPNsense installations that I updated from 16.7.14_2 to 17.7.
- All OPNsense installations are virtual machines running on VMware ESXi 6.5.0.
- All OPNsense installations are configured with 1 CPU core, 1024MB memory, 8GB disk and 3 or 5 E1000 network cards.
- All OPNsense installations are running the plugins for os-vmware and os-snmp
- All VMware servers hardware are HP Proliant DL385p Gen8 with 2 AMD Opteron 6320 processors.
Anyone has any idea what's going on here?
Kind regards,
Bert
-
Hi Bert,
There is a bug in the FreeBSD code of 11.0 at least, causing this, reported at least since December 2016:
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=215471
We migrated to 11.0 in January 2017 with 17.1. There are a couple of threads for this already.
Possible workarounds: load a disk in your VM's CD drive, or remove the CD drive from the VM.
Cheers,
Franco
-
Hi Franco,
You're a hero!
For now, I just mounted an ISO file in the CD drive, and it all works OK now.
Tonight, I will just remove the CD drive completely from the VM config, because I don't need it anyway after initial installation.
Thanks for your quick help.
Kind regards,
Bert
-
Hi Bert,
Sure, here to help. :)
Cheers,
Franco