Recent content by dgeist

P.L., these were my thoughts as well. "Integration" in the true sense of the word is unlikely due to points made above, but happy co-existence and symbiosis seems pretty straightforward as another posix-type storage option for block images. The .apt support is there in the free downloads (just...

Thanks, Udo. I assume that goes in the <VMID>.conf file? I also found that I need serial console access (as opposed to emulated VGA. I was able to make both work properly simply by launching the VMs from CLI, but having it be accessible from the UI would be a boon for management simplicity (part...

Hi, I'm dealing with a commercial kvm-compatible VNF image that I would like to host on a PVE cluster. I've been able to get it to boot on kvm/qemu/libvirt setup on an ubuntu host, but seem to be unable to pass the proper CPU options to the VM under PVE. I would like to do the following: --cpu...

Hi, Tom. I don't think anyone would expect Proxmox or any other purely open source product to adopt support for a closed product. EMC has done a fairly good job with the product as a steward after its purchase making scaleio MORE accessible to the linux community. I've been a FOSS user/advocate...

Given that the ScaleIO system is closed-source and could not be "integrated" into the proxmox environment, there's no reason that it has to be incompatible. I'm looking into trying the same thing based on experiences with CEPH's complexity and the shortfalls of its "one product does everything"...

Yes, just like that, but with two. The latter. I read the man page and wiki :) If you look at the diff, that's what it does. That is the way that every sizeable ceph install I've ever heard of is done (and it's just a good idea to separate the replication from the client mounts). Dan

Actually, nevermind. Here's the diff if you'd like it: >:/usr/share/perl5/PVE/API2# diff Ceph.pm Ceph.pm.orig 687,693c687 < description => "Use specific network for ceph related traffic", < type => 'string', format => 'CIDR', < optional => 1, < maxLength => 128, < }, <...

Hi, after running the ceph setup a few times, I wondered why there was no "cluster network" setup in the install scripts. I found where it's instantiated in the perl module but it just gets the same network address as the "public" address. Granted, this can be edited (and I have) manually, but...

The monitors were so badly out of sync (not time-wise, but otherwise) and I'd tried various things to fix them (exporting maps, editing them, importing, them, etc.). I finally had to call the cluster a loss. I had very little there that wasn't recoverable. I'll need to setup some real-world...

I have a 3 node PVE4.0 cluster running ceph with 3 mons and 18 OSDs (6 per host). I had a situation where all three hosts powercycled at the same time. When they came back, none of the ceph processed would start correctly. Can someone point me to a procedure to get a cluster started correctly...

I've been able to wrestle the PVE permissions system to do most everything I've needed. I'm stuck on how to grant access to ceph, though. As a "pve realm" user that has "Administrator" privileges, I get: Permission check failed (user != root@pam) (403) when attempting to show any ceph attributes...