Ask Your Question
0

libvirtError: internal error: No more available PCI slots

asked 2015-07-20 02:16:43 -0500

prithvee zankat gravatar image

updated 2015-07-20 06:31:24 -0500

When tried to attache a cinder volume to an instance (CentOS release 6.6 (Final) with 26 disks attached), the operation get fails and horizon also don't show any error. Is there any limit on the number of disks that can be attached to a given instance in open stack. Following an error that is collected from nova compute log.

2015-07-20 10:37:37.756 38149 TRACE nova.virt.block_device [instance: 590da105-3cc4-4daa-97a7-1d9c0119f4cc]   File "/usr/lib64/python2.7/site-packages/libvirt.py", line 524, in attachDeviceFlags
2015-07-20 10:37:37.756 38149 TRACE nova.virt.block_device [instance: 590da105-3cc4-4daa-97a7-1d9c0119f4cc]     if ret == -1: raise libvirtError ('virDomainAttachDeviceFlags() failed', dom=self)
2015-07-20 10:37:37.756 38149 TRACE nova.virt.block_device [instance: 590da105-3cc4-4daa-97a7-1d9c0119f4cc] l**ibvirtError: internal error: No more available PCI slots**
2015-07-20 10:37:37.756 38149 TRACE nova.virt.block_device [instance: 590da105-3cc4-4daa-97a7-1d9c0119f4cc]
2015-07-20 10:37:38.347 38149 ERROR nova.compute.manager [req-b79ed15e-b58d-48c1-b144-27fea57b394c f0685bac9133426c85465a9c130a45c1 b5fae37e9190449e93ddfa0215b5166a - - -] [instance: 590da105-3cc4-4daa-97a7-1d9c0119f4cc] Failed to attach 25889774-f351-4a03-ac35-bc8ec99317b8 at /dev/vdab
2015-07-20 10:37:38.347 38149 TRACE nova.compute.manager [instance: 590da105-3cc4-4daa-97a7-1d9c0119f4cc] Traceback (most recent call last):
2015-07-20 10:37:38.347 38149 TRACE nova.compute.manager [instance: 590da105-3cc4-4daa-97a7-1d9c0119f4cc]   File "/usr/lib/python2.7/site-packages/nova/compute/manager.py", line 4778, in _attach_volume
2015-07-20 10:37:38.347 38149 TRACE nova.compute.manager [instance: 590da105-3cc4-4daa-97a7-1d9c0119f4cc]     do_check_attach=False, do_

Thanks, Prithveesingh Zankat

edit retag flag offensive close merge delete

1 answer

Sort by ยป oldest newest most voted
0

answered 2015-07-20 06:21:08 -0500

updated 2015-07-20 06:27:02 -0500

Yes a limit probably exists, and it looks like you've just reached it.

The virtual machine may only have a few virtual pci slots, that are used by the virtual disk controllers.

The error / limit is probably not openstack-specific, it's coming from the underlying virtualization software.

Maybe you could try on bare KVM, without openstack, to see if this limit already exists there. And then try using scsi virtual disk controllers to alleviate the problem, as they should allow for more disks to be connected to each scsi bus. Only if you're not already using scsi controller, though...

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

1 follower

Stats

Asked: 2015-07-20 02:15:52 -0500

Seen: 968 times

Last updated: Jul 20 '15