IPA fails to pull image after devstack node reboot

Bug #1610389 reported by Sukhdev Kapur on 2016-08-05
This bug affects 1 person
Affects Status Importance Assigned to Milestone
Won't Fix

Bug Description

When an instance is launched, the server boots, pulls the ramdisk image successfully. Reboots, IPA tries to pull the image and fails to do so.

This happens with devstack deployment.
This failure occurs only after the openstack controller is rebooted. Unstack/stack clears the issue.

To reproduce this - follow these steps:

1) Do a fresh install of devstack
2) Launch a server - check that it boots fine and attaches to the tenant network
3) Now reboot the openstack controller node
4) repeat step 2 and you will see that IPA will fail to fetch the image. The failure is related to swift timing out the url (and or authentication)

See the following paste for the detailed logs of the error:

Could you please see if a swift disk image is mounted?

$ cat /etc/mtab
/dev/loop2 /opt/stack/data/drives/sdb1 xfs rw,noatime,nodiratime,nobarrier,logbufs=8 0 0

I've heard that it is necessary to mount swift disk image manually after reboot a server built with DevStack.
You can mount the disk image with the following command, which can be seen in /opt/stack/devstack.log.

$ sudo mount -t xfs -o loop,noatime,nodiratime,nobarrier,logbufs=8 /opt/stack/data/swift/drives/images/swift.img /opt/stack/data/swift/drives/sdb1

Sukhdev Kapur (sukhdev-8) wrote :

@Hironori, thanks for your suggestion. Your suggestion corrects the issue. So, it turns out that the reboot of the controller requires remounting to fix the link to the images.

Sam Betts (sambetts) on 2016-09-14
Changed in ironic:
status: New → Won't Fix
summary: - IPA fails to pull image after reboot
+ IPA fails to pull image after devstack node reboot
To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Other bug subscribers