Unable to ignore and proceed when a rolling upgrade task fails.

[Abhey Rana <ab...@gmail.com>]

Aloha Developers,

I was recently upgrading a cluster of 100 nodes using Ambari's rolling
upgrade. During upgrade when almost half of the nodes were upgraded an
upgrade task failed on a particular host due to hard disk failure. In this
case Ambari web UI prompted me to either downgrade or to retry. In this
case retry doesn't seems to be a good option so is downgrade because almost
half of the nodes were upgraded. What could be a better workaround in this
case is, If Ambari could provide me with an option of ignore and proceed as
it provides for service checks, and then stop the service that failed. We
can manage a loss of one host in such a large cluster and can add that host
later with latest bits once the remaining cluster is upgraded.

Has anyone faced a similar issues in past, If yes then please tell me any
probable workaround as of today. Also is there any issue on JIRA regarding
the development of this feature (of ignoring and proceeding) or feature
similar to this, because I was unable to find one.

Please let me know your thoughts on this, because if this problem is being
faced by a lot of orgs then we can create a JIRA issue and then start
development for the same.

Mahalo,
Abhey Rana

CSE Undergraduate | NIT Allahabad

site: abhey.github.io
email: abhey.mnnit@gmail.com
address: Tilak Hostel, MNNIT Allahabad
<https://github.com/Abhey>[image:
https://www.facebook.com/Abhey.Rana.Useless]
<https://www.facebook.com/Abhey.Rana.Useless>
[image: Mailtrack]
<https://mailtrack.io?utm_source=gmail&utm_medium=signature&utm_campaign=signaturevirality5&>
Sender
notified by
Mailtrack
<https://mailtrack.io?utm_source=gmail&utm_medium=signature&utm_campaign=signaturevirality5&>
07/06/18,
11:11:12