HPE 9000 and HPE e3000 Servers
cancel
Showing results for 
Search instead for 
Did you mean: 

Cell failure with Virtual Partitioning in rp8420

 
Sidney Rocha
Occasional Advisor

Cell failure with Virtual Partitioning in rp8420

Hello,

I'm designing a rp8420 with 16 processors. This rp8420 will run 4 vPARs (1 nPAR) with 4 processor each one. I can put 8 processors in two cells (and 2 vPARs per cell board) but for avaiability purposes, I would like to design it using 4 cell boards and one vPAR per cell. My question is: If a failure occurs in one cell, are the other cell boards affected by this failure (it is just one nPAR)? The purpose of this question is to get sure about the availability of the other vPAR in case of failure of a different cell board.

If you have documents with this information avaiable, I would be glad if you can share with me.

Regards,
Sidney Rocha.
3 REPLIES 3
Torsten.
Acclaimed Contributor

Re: Cell failure with Virtual Partitioning in rp8420

A serious hardware failure can affect the whole nPar, in fact all vPars running within the nPar.
Why not creating at least 2 nPars to increase availability?

Hope this helps!
Regards
Torsten.

__________________________________________________
There are only 10 types of people in the world -
those who understand binary, and those who don't.

__________________________________________________
No support by private messages. Please ask the forum!

If you feel this was helpful please click the KUDOS! thumb below!   
Edgar_10
Frequent Advisor

Re: Cell failure with Virtual Partitioning in rp8420

Hi Sidney,

I would tend to agree with Torsten, actually I would tend to advise that since your requirement appears to be that you want 4 environments, I would go for 4 nPARs.

HTH
Sidney Rocha
Occasional Advisor

Re: Cell failure with Virtual Partitioning in rp8420

Hello,

Thank you for your response.

Actually, the customer is interested also in dynamic processor re-allocating. And unfortunatelly we can not do this within nPartitions so far.

What I am surprised is the fact that a cell failure would crash the entire server in a nPar. Are you sure about that? If you have any documents avaiable at HP site or any other place, please let me know. The purpose of the document is making it avaiable for other consultants from my company.

Regards,

Sidney.