Hi,
I've stumbled across a very strange bug and been trying to work around it for a couple of days with no luck.
So my setup is:
Host 1: centos 6.6 with ovirt 3.5 deployed as hosted engine. Had a few problems while deploying it, but eventually got it up and running.
Host 2: same centos 6.6, same hardware. While executing hosted-engine --deploy, I am getting this error:
[ INFO ] Stage: Setup validation
[WARNING] Host name hyp-A2 has no domain suffix
[WARNING] Failed to resolve hyp-A2 using DNS, it can be resolved only locally
[ ERROR ] Failed to execute stage 'Setup validation': [Errno 2] No such file or directory: '/rhev/data-center/mnt/192.168.1.59:_media_ssd__raid10_ovirt_data__fast/f617dfce-4ac8-495b-a0f3-4f85ce39df27/ha_agent/hosted-engine.metadata'
My storage machine is a separate host running RHEL 6.4. /etc/exports looks like this:
/media/ssd_raid10/ *(rw,sync,all_squash)
/media/ssd_raid10/ovirt *(rw,sync,no_subtree_check,anonuid=36,anongid=36,all_squash)
*some other shares*
Now here is the strange part: after repeatedly bashing my head against the wall I finally noticed the difference: on the first host it mounts as:
192.168.1.59:/media/ssd_raid10/ovirt/data_fast on /rhev/data-center/mnt/192.168.1.59:_media_ssd__raid10_ovirt_data__fast_
While on the second host it mounts as:
192.168.1.59:/media/ssd_raid10/ovirt/data_fast on /rhev/data-center/mnt/192.168.1.59:_media_ssd__raid10_ovirt_data__fast
(note the lack of underscore in the end)
That is what caused the problem - this "hosted-engine.metadata" file is actually a symlink. And because of different mount points on the second host it was pointing to nowhere.
So I figured, maybe I should add / to the end of the nfs path. So I tried it like this:
Please specify the storage you would like to use (iscsi, nfs3, nfs4)[nfs3]:
Please specify the full shared storage connection path to use (example: host:/path): 192.168.1.59:/media/ssd_raid10/ovirt/data_fast/
[ INFO ] Installing on first host
Please provide storage domain name. [hosted_storage]:
As you can see, in this case it doesn't detect that it already has a hosted engine installed.
So, I'm out of ideas how to make it work. There is clearly something unstable in the installation scripts, but does anyone have any idea how to make it work?
PS: Sorry for my crude English, not really a native speaker.
Thanks,
Tigran.