Hello, On Tue, Dec 27, 2022 at 8:40 AM Yedidyah Bar David <didi@redhat.com> wrote:
Sorry, I do not follow. Is your immediate obstacle being that engine-setup refuses to continue, saying "Hosted Engine HA is in Global Maintenance mode."?
You can cause it to ignore this test by passing 'OVESETUP_CONFIG/continueSetupOnHEVM=bool:True' (in the answer file or --otopi-environment).
We recently added an option 'engine-setup --show-environment-documentation', exactly for this env key, see also:
https://bugzilla.redhat.com/show_bug.ccontinueSetupOnHEVM=bool:Truegi?id=170... <https://bugzilla.redhat.com/show_bug.cgi?id=1700460>
Best regards, -- Didi
I actually managed to bypass the check by editing he.py and deleting the "raise" statement, preventing hosted-engine from bombing out because it wasn't able to connect to the nodes. From there I managed to renew the certificates (see second mail), and even connected two of the 3 nodes successfully (I had to create new temporary vdsm certificates, get them semi-connected to the engine, and then "re-enroll certificates" from the UI. Once I had a limping cluster up, I shut everything down cleanly, and... and redeployed the cluster from scratch. (with all the failed attempts, my HE was completely busted). That said, I wonder if having to short circuit the environment variable isn't a bit over-complicated, given the considerable number of cert related issues. But thanks for the heads-up. Q: I'm willing to try and document all the steps I did, in my semi-success attempt to save my cluster. That said, I rather not document wrong / broken steps. Can anyone @RH review my writeup? - Gilboa