Posts made by Heady

@Heady

Update - I have been able to make progress.

(1) I have been able to get XO version 5.80.0 appliance installed. For some reason the automated web deployment method didn't work. Couldn't work out why.
(2) Had to reinstall xcp-ng again on the new host.
(3) Using XO appliance - had to add a number of interfaces prior to attempting to add to the pool.
(3) Using XO appliance - have been able to add the new host to the pool.
(4) Took a few reboots and manual powercycles to get the new host integrated into the pool and recognized within XO appliance. Shutdown/reboot would hang and not complete.
(5) The SRs had to be connected / disconnected / connected from the XO Host-Storage tab for them to be online and accessible. Couldn't use connect to all hosts from the Pool-Storage tab.
(6) Then finally success in moving VMs from the old host to the new host.

Phew... half-way there.

Seem to be having trouble adding new host to pool...

Tale of woe: because of the storms - had a host within a 2 host pool die (the HDD gave-up the ghost). And I'm now worried that the HDD in the remaining host being the same age may fail in future. So was going to install new HDDs in both hosts while keeping the pool going.

My plan was to:
(1) Change the dead HDD in the affected host.
(2) Install same XCPng version already running (ver 8.0.0)
(3) Eject dead host from pool & clean-up.
(4) Then add newly installed host to existing pool.
(5) Move all VMs to new host.
(6) Designate new pool master.
(7) Eject old host from pool.
(8) Power-down old host & replace HDD.
(9) Install same XCPng version.
(10) Add old host to pool.
(11) Follow upgrade instructions from docs 8.0.0->8.2.0
(12) Sigh in relief.

Steps (1) to (3) were successful. However, my plan has failed at step (4).

Could not add new host to pool via XO (old ver. 5.31.1) as it would error with: "getaddrinfo EAI_AGAIN updates.xensource.com:80".

So I added new host to pool via xe but it has only done half the work. It has configured the management bond0 interface but not the storage interface - and has not completed the pool vlans etc... It has database entries for the SRs - but they remain unplugged as unavailable. Via xe the new host is seen in the pool. The VM list can been seen on both hosts but because the SRs and vlans are not synchronized cannot migrate VMs.

This half-added-to-pool-state has caused issues with XO. I now have log errors of "Internal Error", "server already exists", "no such pool", "the server's pool is already connected" ...

Any ideas of where to go from here?

(1) Stop XO being confused over the new host?
(2) Sync everything correctly between new host and current pool master?

Thx