Ask Your Question
0

instance cannot attach volume, although compute-host establishes valid session?

asked 2013-05-21 03:56:58 -0500

digitalwonk gravatar image

I am using grizzly stable from cloud archive, and currently have a problem with cinder volumes not being mounted within instances. I am using tgtadm as the iscsi helper, which is set on cinder, nova controller, and nova compute hosts. The iscsi targets seem to be connected within the compute-host, as seen by iscsiadm. Is there a particular config option or setting that might cause this behavior? Or is there a way I can further investigate what the problem might be?

This appears to be the only warning I receive on the compute host:

/var/log/nova/nova-compute.log 2013-05-20 20:32:19.882 AUDIT nova.compute.manager [req-fbb43291-5251-4514-b977-8ce3cbbdfaf1 a0c94431c8f942b9a8b67d7cab818364 e953f96c7b2a4598ae0b507b936970e2] [instance: 8f530429-7a86-49e2-b5d8-505a75d7669d] Attaching volume e48ecedb-8580-4262-8488-97f8d7c116ee to /dev/vdc 2013-05-20 20:32:22.021 WARNING nova.virt.libvirt.volume [req-fbb43291-5251-4514-b977-8ce3cbbdfaf1 a0c94431c8f942b9a8b67d7cab818364 e953f96c7b2a4598ae0b507b936970e2] ISCSI volume not yet found at: vdc. Will rescan & retry. Try number: 0

From the /var/log/syslog, the compute host seems to have mounted it correctly: May 20 20:32:20 os-compute04 kernel: [24493.934883] scsi11 : iSCSI Initiator over TCP/IP May 20 20:32:21 os-compute04 kernel: [24494.945033] scsi 11:0:0:0: Direct-Access IET VIRTUAL-DISK 0 PQ: 0 ANSI: 4 May 20 20:32:21 os-compute04 kernel: [24494.945495] sd 11:0:0:0: Attached scsi generic sg3 type 0 May 20 20:32:21 os-compute04 kernel: [24494.948319] sd 11:0:0:0: [sdc] 62914560 512-byte logical blocks: (32.2 GB/30.0 GiB) May 20 20:32:21 os-compute04 kernel: [24494.949132] sd 11:0:0:0: [sdc] Write Protect is off May 20 20:32:21 os-compute04 kernel: [24494.949139] sd 11:0:0:0: [sdc] Mode Sense: 77 00 00 08 May 20 20:32:21 os-compute04 kernel: [24494.950632] sd 11:0:0:0: [sdc] Write cache: disabled, read cache: enabled, doesn't support DPO or FUA May 20 20:32:21 os-compute04 kernel: [24494.957304] sdc: unknown partition table May 20 20:32:21 os-compute04 kernel: [24494.960315] sd 11:0:0:0: [sdc] Attached SCSI disk May 20 20:32:21 os-compute04 dnsmasq-dhcp[2716]: DHCP packet received on br-ex which has no address May 20 20:32:22 os-compute04 iscsid: connection1:0 is operational now

When I perform the "iscsiadm -m session -o show", I get this session, which is the correct volume tcp: [1] 10.0.0.137:3260,1 iqn.2010-10.org.openstack:volume-e48ecedb-8580-4262-8488-97f8d7c116ee

However, when I login to the VM, the device is not created (or any new device for that matter). Thanks in advance for any assistance.

edit retag flag offensive close merge delete

4 answers

Sort by ยป oldest newest most voted
0

answered 2013-05-23 18:37:07 -0500

digitalwonk gravatar image

It needs to be loaded as part of an image or within an instance before you attach a cinder volume.

edit flag offensive delete link more

Comments

what's "be loaded" mean?

wangxh gravatar imagewangxh ( 2014-04-12 22:48:02 -0500 )edit
0

answered 2013-05-23 06:58:15 -0500

shantha-kumar gravatar image

where you need to install the acpiphp module ?

Is it inside the launched instances or in compute node ?

edit flag offensive delete link more
0

answered 2013-05-21 14:34:20 -0500

digitalwonk gravatar image

It turned out to be the acpiphp module that was missing from within the vm, at least for one of my images.

edit flag offensive delete link more
0

answered 2013-05-21 13:22:28 -0500

digitalwonk gravatar image

On another vm, I attempted to attach a volume with the same results. I was able to determine that the /etc/libvirt/qemu/instance-xxxx.xml, the following was created:

<disk type='block' device='disk'>
  <driver name='qemu' type='raw' cache='none'/>
  <source dev='/dev/disk/by-path/ip-10.0.0.137:3260-iscsi-iqn.2010-10.org.openstack:volume-f30aa828-c1dd-4786-9cde-7d36f6be9777-lun-1'/>
  <target dev='vdd' bus='virtio'/>
  <serial>f30aa828-c1dd-4786-9cde-7d36f6be9777</serial>
  <address type='pci' domain='0x0000' bus='0x00' slot='0x07' function='0x0'/>
</disk>

Hopefully, that will provide a clue.

edit flag offensive delete link more

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Get to know Ask OpenStack

Resources for moderators

Question Tools

2 followers

Stats

Asked: 2013-05-21 03:56:58 -0500

Seen: 158 times

Last updated: May 23 '13