Hello,
I'd install proxmox v5 with HA + Fibre Channel multipath. For now there are two nodes (shoulde be 3), but third node will be added in this week.
My target is:
shared storage for containers (bind mounts)
Containers should be failover (one dedicated LUN) based on proxmox storage model LVM with locking pve mechanism and it's ok.
Why:
For example I have php7.0 container on node1, php5.6 on node2 and php7.1 on node3. All of them have apache2.
I don't want migrate folders (with php code and data) from php5.6 to php7.0.
I would like only move virtualhosts config to apache on other container.
With bind mounts I cane share for example WWW folder with php code and data to all containers on diffrent nodes.
I don't have big storage and I can't always migrate with rsync specified directory contain WWW files between containers and nodes.
With this solution I don't need planned how many storage I need for php5.6, 7.0 or 7.1, and then resize it or worst reduce.
LUN created on SAN (Fibre Channel) should be available and accessible from all nodes (done)
Now I wondering about share it filesystem (gfs2) on logical volume to all node.
Description:
I installed DLM (dlm-controld), created LVM (not clvm) and format it with gfs2.
On first node this LVM is active, and on the others I had to activate manually.
I had one problem after freshly created gfs2, reboot one node1 cause inacessible filesystem on the others.
After reboot all of them situasion seem be ok.
Now reboot one of them not affect to the rest. Maybe reload pve cluster would be enough.
I added entry to fstab based on UUID, after reboot node I must mount manually filesystem with command 'mount -a', why?
I can't created CLVM because of error message which I found on the forum, but without responses, because purposes were diffrent.
There is no systemctl clvmd service, only command clvmd.
I heard that gfs2 should be on LVM, not on raw block disk.
Questions:
Why LVM filesystem isn't mounted after reboot?
Can I stay with LVM (not CLVM) with gfs2, is it safe?
Maybe there is another solution?
Edit: 2017-09-18
CLVM need :
systemctl enable lvm2-cluster-activation
locking_type 3 in /etc/lvm/lvm.conf
vgchange -c y VOLUME_GROUP
We can create script in /etc/init.d to mount gfs2 lvm as directory. GFS2 must be umount during reboot or shutdown (umount -a -t gfs2 or leave entry in fstab).
best regards,
nonton
I'd install proxmox v5 with HA + Fibre Channel multipath. For now there are two nodes (shoulde be 3), but third node will be added in this week.
My target is:
shared storage for containers (bind mounts)
Containers should be failover (one dedicated LUN) based on proxmox storage model LVM with locking pve mechanism and it's ok.
Why:
For example I have php7.0 container on node1, php5.6 on node2 and php7.1 on node3. All of them have apache2.
I don't want migrate folders (with php code and data) from php5.6 to php7.0.
I would like only move virtualhosts config to apache on other container.
With bind mounts I cane share for example WWW folder with php code and data to all containers on diffrent nodes.
I don't have big storage and I can't always migrate with rsync specified directory contain WWW files between containers and nodes.
With this solution I don't need planned how many storage I need for php5.6, 7.0 or 7.1, and then resize it or worst reduce.
LUN created on SAN (Fibre Channel) should be available and accessible from all nodes (done)
Now I wondering about share it filesystem (gfs2) on logical volume to all node.
Description:
I installed DLM (dlm-controld), created LVM (not clvm) and format it with gfs2.
On first node this LVM is active, and on the others I had to activate manually.
I had one problem after freshly created gfs2, reboot one node1 cause inacessible filesystem on the others.
Code:
commands df -h, ls, dd, touch etc. not working and hang up bashNow reboot one of them not affect to the rest. Maybe reload pve cluster would be enough.
I added entry to fstab based on UUID, after reboot node I must mount manually filesystem with command 'mount -a', why?
Code:
UUID=6fabd54e-184d-2155-1de6-a9f02f1a937f /mnt/WWW gfs2 noatime,_netdev 0 0I can't created CLVM because of error message which I found on the forum, but without responses, because purposes were diffrent.
Code:
connect() failed on local socket: No such file or directory
Internal cluster locking initialisation failed.
WARNING: Falling back to local file-based locking.
Volume Groups with the clustered attribute will be inaccessible.There is no systemctl clvmd service, only command clvmd.
I heard that gfs2 should be on LVM, not on raw block disk.
Questions:
Why LVM filesystem isn't mounted after reboot?
Can I stay with LVM (not CLVM) with gfs2, is it safe?
Maybe there is another solution?
Edit: 2017-09-18
CLVM need :
systemctl enable lvm2-cluster-activation
locking_type 3 in /etc/lvm/lvm.conf
vgchange -c y VOLUME_GROUP
We can create script in /etc/init.d to mount gfs2 lvm as directory. GFS2 must be umount during reboot or shutdown (umount -a -t gfs2 or leave entry in fstab).
best regards,
nonton
Last edited: