Unsolved
This post is more than 5 years old
1 Rookie
•
12 Posts
0
3525
August 21st, 2012 08:00
Utility Node Failure
Hi All
Want to get clarity on utility node since it is a single point of failure in avamar RAIN setup. If a utility node is completely destroyed, how much time would it take to resume the function of the utility node on a spare node and the process to accomplish the same as compared to getting a new node shipped.
Also which of the two is a better option, promoting a spare node to a utility node or replacing the utility node with a newly shipped one and why?
TA
No Events found!
ionthegeek
2 Intern
•
2K Posts
0
August 21st, 2012 09:00
The process for rebuilding a destroyed utility node is the same regardless of the type of node being used as the target (new utility node or spare node).
The utility node replacement process is documented in the following technical notes (available through PowerLink or Support Center):
300-007-551 - AVAMAR UTILITY NODE REPLACEMENT PROCEDURE (Avamar v5.x and earlier)
300-012-538 - UTILITY NODE REPLACEMENT PROCEDURE (Avamar v6.0)
Edit: The most current procedures can now be found in SolVe Desktop.
I would strongly recommend engaging support if you lose a utility node. The rebuild process is non-trivial. I would block out an 8 hour window for the rebuild but an experienced support person can complete the process considerably faster.
I would personally recommend having a replacement utility node shipped out to the site rather than using the spare. Using a spare node as a utility node is supported but it is not a typical configuration on Gen2 or newer hardware. Sticking to a standard configuration wherever possible makes life easier for support which will, in turn, make life easier for you.
Druehl1
2 Intern
•
223 Posts
0
April 22nd, 2014 04:00
Hi Ian,
has there been any changes in replacing a utility node in version 7?
Is there a document for avamar customers, where there is described what happend if a utility node fails and what can be done if this happens.
For example if I have 2 grids, can I replicate the configuration to the a secound grid and restore the configuration and go on?
If there are hardware errors, can only the disks can be replacned in a new node, and then go on?
If there are in the worst case disk errors, so that the utility node don´t come up again, what is the process here?
I dont need any detailed technical solution for this, it is only to describe the customer this process.
ionthegeek
2 Intern
•
2K Posts
1
April 22nd, 2014 07:00
The process is largely unchanged for version 7. The only document I'm aware of is the full replacement procedure but that covers all the technical detail.
I would not recommend trying to perform any kind of configuration replication except as described in the utility node replacement procedure (which you can now find in SolVe Desktop, the successor to the Avamar Procedure Generator).
There are two replacement procedures that can be used depending on the nature of the failure. Both can be found in SolVe desktop.
If the RAID set is intact, the utility node can be replaced using the disk transplant / chassis swap procedure. There are a few extra steps for utility nodes since the license is keyed on the utility node hardware and has to be regenerated. Naturally, you can only transplant the disks into a utility node chassis.
If the RAID set is destroyed or the file system is otherwise unusable, a full utility node rebuild is required. The general outline of the procedure is:
Druehl1
2 Intern
•
223 Posts
0
April 22nd, 2014 07:00
Hi Ian,
thanks for your answer.
One last question.
in case of a RAID set failure, would it be good to make a backup of the probe.xml files and network config to another system?
ionthegeek
2 Intern
•
2K Posts
0
April 22nd, 2014 09:00
It couldn't hurt to tar up the probe.xml and networking files and stash them away somewhere safe.