Guest Install Issues on Proxmox 4.0

[adamb]

Started some testing with Proxmox 4.0 as it looks pretty sweet. Every was going fantastic until I went to install a CentOS 6.6 VM. I am hitting this exact issue except I don't run a Dell array, its a Nimble iSCSI array.

https://access.redhat.com/solutions/1175843

Using anything else other than cache=none works and the VM installs. Trying to determine if this is a known bug/issue. Has anyone else run into this issue?

 

[t.lamprecht]

Can you try their advice to install the guest onto a device whose logical block size is 512 bytes rather than 4K?

Other VMs are working fine?

 

[adamb]

Can you try their advice to install the guest onto a device whose logical block size is 512 bytes rather than 4K?

Other VMs are working fine?

The storage solution I am using as far as I know doesn't do fixed block sizes. It is a variable block size architecture.

CentOS6 and CentOS7 are both having the issue. I haven't tried another OS as of yet. I will have to try another OS and report back.

 

[adamb]

Found a bit more detail. The install only bombs out when using a disk type of virtio and cache=none. If I use any other disk type I can use cache=none. The only way I can use a disk type of virtio is to not use cache=none.

OS's tested so far
CentOS 5
CentOS 6
CentOS 7
Windows 7

I am thinking this is a bug and I should create a bugzilla. Let me know your thoughts.

 

[adamb]

This issue just got more interesting. I can only recreate the issue with the first VM. All VM's after the 1st one install with no issues. I can remove the first VM, then re-add and the issue persists. This is really odd!

 

[t.lamprecht]

Okay this sound really odd and like the results from a lot coincidences and special setup,

The other VMs are exactly the same like the first one (config wise)?
What if you add two disks to the first vm and select the second?

 

[adamb]

Okay this sound really odd and like the results from a lot coincidences and special setup,

The other VMs are exactly the same like the first one (config wise)?
What if you add two disks to the first vm and select the second?

All other VM's are identical config wise. I added a 2nd disk to the 1st VM and it has no issues installing. The issue only persists on the first disk of the first VM.

This is a standard 3 node setup using HP DL380p Gen9's with a Nible iSCSI C300 array.

 

[t.lamprecht]

And did you try to install another distribution on the first disk?
If that also fails, it may be a problem with your iSCSI array (fishing a bit in the dark here, odd issue).

 

[adamb]

And did you try to install another distribution on the first disk?
If that also fails, it may be a problem with your iSCSI array (fishing a bit in the dark here, odd issue).

Yep all other OS's fail on the first disk. I do agree the issue lends itself to the array. I appreciate the input.

 

[adamb]

Ok, bringing this issue back up. After further testing I found that the VM and the Host would take filesystem errors like the one below. This happens in a C6 VM, a C7 VM and directly on the host when I mount the block device formated in ext4. I thought for sure this ment it was something with the iSCSI array. As a last test, I loaded up a fresh proxmox 3.4 and the issue is completely gone. Not only is the filesystem issue gone, but so is the original issue of not being able to install to the first created disk. If I go right back to my proxmox 4.0 host both issues are 100% reproducible. I currently have a proxmox 3.4 host and 4.0 host setup on identical HP DL 380p gen 9's. Both of them are connected to the same iscsi storage sharing the same LVM space. I can move the VM config between the two and reproduce the issue 100% of the time on proxmox 4.0, this does not happen on 3.4.

I guess I will just leave Proxmox 4.0 alone until its officially released. I was hoping to move forward with it in testing and get a good feeling! Luckily this is only for inhouse but we do see a need to move to proxmox 4.0 for our customers at some point as I am pretty sure wheezy only has a few more years left of support.

Sep 16 07:51:28 linux151 kernel: EXT4-fs error (device vda4): ext4_mb_generate_buddy: EXT4-fs: group 108: 268 blocks in bitmap, 32768 in gd
Sep 16 07:51:28 linux151 kernel: EXT4-fs error (device vda4): ext4_mb_generate_buddy: EXT4-fs: group 109: 943 blocks in bitmap, 32768 in gd
Sep 16 07:51:28 linux151 kernel: EXT4-fs error (device vda4): ext4_mb_generate_buddy: EXT4-fs: group 110: 0 blocks in bitmap, 32768 in gd
Sep 16 07:51:28 linux151 kernel: EXT4-fs error (device vda4): ext4_mb_generate_buddy: EXT4-fs: group 111: 0 blocks in bitmap, 32768 in gd
Sep 16 07:51:28 linux151 kernel: EXT4-fs error (device vda4): ext4_mb_generate_buddy: EXT4-fs: group 113: 0 blocks in bitmap, 32768 in gd

 

[adamb]

Forgot to mention the current packages I am running.

root@ccsmiscrit2:~# pveversion -v
proxmox-ve: 4.0-10 (running kernel: 4.2.0-1-pve)
pve-manager: 4.0-36 (running version: 4.0-36/9815097f)
pve-kernel-3.19.8-1-pve: 3.19.8-3
pve-kernel-4.1.3-1-pve: 4.1.3-7
pve-kernel-4.2.0-1-pve: 4.2.0-10
lvm2: 2.02.116-pve1
corosync-pve: 2.3.4-2
libqb0: 0.17.1-3
pve-cluster: 4.0-17
qemu-server: 4.0-23
pve-firmware: 1.1-7
libpve-common-perl: 4.0-20
libpve-access-control: 4.0-8
libpve-storage-perl: 4.0-21
pve-libspice-server1: 0.12.5-1
vncterm: 1.2-1
pve-qemu-kvm: 2.4-5
pve-container: 0.9-18
pve-firewall: 2.0-11
pve-ha-manager: 1.0-5
ksm-control-daemon: 1.2-1
glusterfs-client: 3.5.2-2+deb8u1
lxc-pve: 1.1.3-1
lxcfs: 0.9-pve2
cgmanager: 0.37-pve2
criu: 1.6.0-1
zfsutils: 0.6.4-pve3~jessie

 

[adamb]

I moved completely back to 3.4 for testing and all of my issues are completely gone. I have filed a bug on the issue. If I need to I can throw in a 4.0 host to do testing when/if you guys fix this bug.

https://bugzilla.proxmox.com/show_bug.cgi?id=724