Recent content by hacman

This is probably not the answer you're looking for, but I just bit the bullet and upgraded the cluster to Squid. It has been ok in testing and has been out for 2 months now without any serious issues so I figured I would just run with it!

Hi all, Having a bit of an odd one here. I've just rebuilt a node in my cluster, installing from scratch using the latest install media. When trying to install Ceph Reef, the following is given. It seems ceph-fuse and ceph-common have already been installed from the squid repo? Reading...

Same issue - SAS disks show unknown, SATA shows as expected. Both types of disks connected the same way.

CloudLinux who develop KernelCare and other products are very targeted in where they spend their development resources. It is not the first time they have added support for software on their products only to quietly let all development efforts languish. Raise a ticket with them and ask what...

Hi all, Shortly we will have to do some work to all our nodes, whereby we will re-create the boot device/partition (switching from hardware RAID to HBA with ZFS). Ideally, it would be nice to not have to go through the whole process of kicking the node out of the cluster, and then fully...

Following up here as I had the same issue. There was a setting in my BIOS/UEFI called C.A.M. (Clever Access Memory). It is related to Resizable BAR, which is mainly used if you're doing gaming or video stuff. Disabling this setting solves the issue for me.

Just upgraded my home machine (Microserver Gen8) in about 15 minutes end-to-end, with absolutely no issues. Very smooth so far! Excellent job guys!

Hi, Separate network which is used for accessing the management-UI and corosync only. No duplicate addresses, no duplicate hostnames. There is no packet loss and good response times between the hosts. This was the first thing I checked, and part of why this issue is so perplexing. There is...

Hi, Nope - I've not changed any of the settings, so it is disabled at the datacenter level. I've just disabled at the node level too so we can be sure, and no difference.

Ok, this had the effect I sort of expected. They clustered fine - and now show up in eachothers web-manager. Output of the cluster command returned: root@NODE-2:~# pvecm add 10.100.2.1 --use_ssh root@10.100.2.1's password: No cluster network links passed explicitly, fallback to local node IP...

I've considered trying that, but was unsure about doing so as I believe once the nodes are clustered they will still need the API to be able to work effectively. I will try it now.

SSH from one host to the next works fine. No hardening or security measures applied - it's a fresh, out the box, just updated install on both machines.

Hi all, We're in the process of setting up a new cluster, with Proxmox 6.4, but have an issue where for some unknown reason we are unable to have the hosts join the cluster. To give some context of the setup, there are currently 2 nodes loaded up, each has had all the latest no-sub updates...

This one was just for Corosync, as part of the pre-upgrade steps. In the end we've just rebuilt the node. Thanks everyone for the help anyway!

Hi Aaron, Sadly that just does the same: root@[HOST]:~# apt --fix-broken install Reading package lists... Done Building dependency tree Reading state information... Done Correcting dependencies... failed. The following packages have unmet dependencies: ceph-common : Depends: libcurl4...