You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
On bare-metal deployments, one gets quickly in a situation where only a full Ambari reset incl. uninstall helps to fix a broken cluster deployment (p.eg we just had applied a cluster blueprint, where the components installation failed, because of bad repository configs).
Of course there's a documented process for doing this manually: https://community.hortonworks.com/questions/1110/how-to-completely-remove-uninstall-ambari-and-hdp.html
..but as ansible (power)users, manual isn't the way and these steps are easy to put in a playbook.
( Handling the DB cleaning generically for the role, might be more work to support postgres, mysql, mariadb . For this we should somehow re-use the distro specific playbooks/roles/database/vars)
Motivation: I created the issue, to get inputs from the community if maybe someone already started work to use ansible for uninstalling.. and if not, I will start with a playbook for 80% and share it via a PR (of course I'ld glad to contribute to this repository)
ps: Ref. to older issue, requiring uninstalling: #53 (comment)
Other than that... no idea, something must be messed up in your ambari. Try to clean it as much as you can (yum remove, rm -rf /var/lib/ambari* /usr/lib/ambari*, drop the ambari database, etc).
The text was updated successfully, but these errors were encountered:
I can imagine this could be going on many levels:
i) your Ambari is good but you messed something in the blueprint -> stop ambari and drop the db, no need to remove ambari
ii) remove all traces of ambari, including the repo (the chosen ambari version is wrong)
iii) remove ambari and hdp/hdf, etc
But ultimately, the hope is that this shouldn't be needed that frequently...
On any bare-metal installs, there should be some testing done on a temporary environment (either public or private cloud) and once all the variables and/or playbook modifications are done, these can just be replayed on the on-prem so it's installed from the first attempt.
I have something that I hacked together that I use for development purposes of uninstalling and reinstalling ambari without rebuilding my whole environment. it does an ambari-server reset and then removes rpms and directories. This only works with default locations currently.
On bare-metal deployments, one gets quickly in a situation where only a full Ambari reset incl. uninstall helps to fix a broken cluster deployment (p.eg we just had applied a cluster blueprint, where the components installation failed, because of bad repository configs).
Of course there's a documented process for doing this manually:
https://community.hortonworks.com/questions/1110/how-to-completely-remove-uninstall-ambari-and-hdp.html
..but as ansible (power)users, manual isn't the way and these steps are easy to put in a playbook.
( Handling the DB cleaning generically for the role, might be more work to support postgres, mysql, mariadb . For this we should somehow re-use the distro specific
playbooks/roles/database/vars
)Motivation: I created the issue, to get inputs from the community if maybe someone already started work to use ansible for uninstalling.. and if not, I will start with a playbook for 80% and share it via a PR (of course I'ld glad to contribute to this repository)
ps: Ref. to older issue, requiring uninstalling:
#53 (comment)
The text was updated successfully, but these errors were encountered: