I setup a GPFS v18.104.22.168 cluster between 2xAIX v5.3.0-TL12-SP5 power740 LPARs.
One GPFS file system was created on 3xEMC SAN directly attached disks configured as NSD devices.
I ran successfull cluster node failover tests and saw that the time for safe recovery of the file system on the second cluster node was measured to 21sec.
Client is going to run his BSCS billing application on that file system.
Client is asking what is the impact-feeling that the appl. users will receive upon cluster node failures during that time?
Will they just see some kind of IO suspension on their appl. gui untill the file system is safely recovered on the second cluster node?
THX in advance.
Pinned topic Application impact after GPFS v3.4 node failure
Answered question This question has been answered.
Unanswered question This question has not been answered yet.
Updated on 2012-10-17T21:35:43Z at 2012-10-17T21:35:43Z by ezhong
ezhong 060001JWNJ33 Posts
Re: Application impact after GPFS v3.4 node failure2012-10-17T21:35:43ZThis is the accepted answer. This is the accepted answer.An application accesses a filesystem through a mount point. If the mount point remains accessible during the cluster node failover, then the application should not be affected. Otherwise, it would depend on how the application handles a filesystem failure and how the mount point get reestablished after failure.
This article has some good pointers on how to keep a GPFS cluster up and running: