Issues with new cluster and migration.

B

burritoincognito

Gawd
Joined
Sep 17, 2012
Messages
767
I've got VCSA 6.7 set up on a Precision 3620 with local SSD storage. I was trying to update, but can't put the machine in maintenance mode while the VCSA is online. So, I picked up a T1650 to add to my home lab. It's an Ivy Bridge with half the RAM, and only a single 256GB SSD. More than enough for me to push a couple systems to when I'm messing around.

Unfortunately, while I was able to create a cluster, I can't join the 3620 to it while it's online, and I can't migrate the VCSA to the 1650 because of the disparity on hardware. I'd need to adjust the cluster to get them compatible with each other by setting it to "max CPU Ivy Bridge" or something.



How do I make these changes while the VCSA is running things, when it needs the machine to be off to make the changes? I'm almost to the point of deciding to just scrap the VCSA and build a new one on the T1650 to manage the network, and hope that it has less problem moving up than moving down.
 
So, the original thing in this post never fixed it...but since this is still front page and relevant, I'll just expand on this rather than posting a new one.

I'm shifting the 1650 to gaming. I wasn't using it since clustering never really worked properly. The 3620 is going back to the employer, since it's their PC checked out for home use.

I'm buying a pair of NUC8i5BEH. I've used the 4th, 6th, and 7th generation NUCs in my home lab before as solitary systems, and I've got enough stuff sitting around to make them both have quad cores, a single 16GB stick of RAM, and 1TB to 1.5TB of SSD space. Identical hardware will mean I should be able to migrate between systems properly...but is there any good reason I should rebuild the VCSA, other than the fact that I'll be able to upgrade from 6.7 to 6.7u1 while doing so? With the issues I ran into, I don't know if the original setup being built on Skylake now sitting on Coffee Lake hardware will even matter.
 
Sounds like you need to turn on EVC on the cluster so it uses the features of the older processors across the board.
Then you could add/remove the hosts as you see fit (assuming the VMs are off).
 
burritoincognito said:
So, the original thing in this post never fixed it...but since this is still front page and relevant, I'll just expand on this rather than posting a new one.

I'm shifting the 1650 to gaming. I wasn't using it since clustering never really worked properly. The 3620 is going back to the employer, since it's their PC checked out for home use.

I'm buying a pair of NUC8i5BEH. I've used the 4th, 6th, and 7th generation NUCs in my home lab before as solitary systems, and I've got enough stuff sitting around to make them both have quad cores, a single 16GB stick of RAM, and 1TB to 1.5TB of SSD space. Identical hardware will mean I should be able to migrate between systems properly...but is there any good reason I should rebuild the VCSA, other than the fact that I'll be able to upgrade from 6.7 to 6.7u1 while doing so? With the issues I ran into, I don't know if the original setup being built on Skylake now sitting on Coffee Lake hardware will even matter.
Click to expand...

Read release notes and see if anything you want to test/use sticks out as fixed/updated in 6.7 U1.

you can always do in place upgrade also from 6.7 to 6.7 U1
 
You must log in or register to reply here.
Back
Top