Search results

I'm getting the error below after something happened (it was not happening before) and not sure that I changed anything deliberately. It prevent the status graphs (rrd, right?) to be displayed on the PBS administration section. Oct 11 22:07:17 pbs3 systemd[1]: Starting...

Thanks for this, @Chris! I have implemented a local PBS and will be running sync jobs to pull the backups off-site. I see how it goes in the next few days.

I have run into an issue a couple of times in that guest OS's slow down dramatically if the PBS server doesn't perform for whatever reason. Previously I had a network issue, which prevented backups from being written at a reasonable speed and it caused the guest machines being backed up to...

From the pfSense on NodeB, I can ping a VM on nodeC on VLAN12, which then mean it is actually working! There is not other way to that a ping from 192.168.161.253 (pfSense VLAN12 gateway on running on NodeB) could reach a VM on NodeC! Since I have only done the changes for Node B and C, testing...

Maybe there's a problem with these SDN ports in pfSense. I have 2 pfSense VM's that act as router firewalls. The default gateway on each network an SDN interface on these pfSense machines. It's actuall a CARP ip address that "floats" between the two firewalls and redirects to the actual...

NodeC:

I had to rename them with .txt extension to be able to attach them here. These are NodeB, NodeC to follow...

Just a bump... @spirit: The SDN is still not working despite the change to a name that is recognized by the code...

I changed two nodes with this setting last night and named the ports enlan0, enlan1, enlan2 and enlan3. However, after this change the guest on one of the nodes still cannot ping a guest on the other node via a SDN VLAN. I did change the /etc/network/interfaces file and rebooted each node. #...

Thanks for checking! Could we do one of two things then please? 1. Create a note in the SDN documentation explaining the limits of nic names used for SDN bridges 2. Allow more options, like lan*, nic*, net* and possibly others Or both? For now, I'm going to name the nics enlan0, enlan1...

I'll have to schedule to test this carefully, since I cannot risk taking down a node (like I did before) and then it doesn't come back up because the ports have different names than I expected. I'd like to report this as a bug, but I'm not sure what is the cause of this. Is it the systemd...

More... VLAN12 is an SDN and vmbr2 is a manually configured bridge to vlan 25. root@FT1-NodeA:~# ethtool VLAN12 Settings for VLAN12: Supported ports: [ ] Supported link modes: Not reported Supported pause frame use: No Supports auto-negotiation: No Supported FEC modes...

I see these in the syslog of the nodes. On the nodeA when the standby firewall is starting: Sep 19 17:14:00 FT1-NodeA systemd-udevd[3462078]: ethtool: autonegotiation is unset or enabled, the speed and duplex are not writable. Sep 19 17:14:01 FT1-NodeA kernel: [564541.374585] device tap101i6...

I wrote in my OP: Therefor I can't use ethX, the system has the ability to use those names automatically, so it messed the nics up by swapping them randomly.

I will restart a node after hours and see if that makes any difference.

I completely removed VLAN12. I first removed the interfaces from the firewall machines, then I removed the SDN config. I then restarted the firewall, recreated the SDN for VLAN12, then re-added the interface (which is a bridge on vmbr0), recreated the rules on the pfSense after adding the ip...

Yes, before I did that, the SDN's showed an error. After applying the status returned to normal. Yes, I have actually deleted an complete definition and recreated it, but it has not no noticeable effect. I think I'll just do one again.

Surely someone much have some idea of what is wrong?

32 nodes with 2 x AMD 128core Epyc V4 processors = 8192 vCPU's per cluster. Pretty big. But as others have said, the inter-cluster scripts and automations are the key to making it work.