We're planning on migrating our clients from the 3.4 series to the 3.5 series, but for various reasons too irritating to go into, we'd like to wait a bit before moving the servers to the new release series.
We know (through a previous PMR) that we need to set scatterBuffers=no for the 3.5 clients if they'll be dealing with 3.4 servers, but is anyone aware of any other interoperability issues that might come up? I've been through the changelogs pretty thoroughly, but they're dense, so it's entirely likely I've missed something. Here's what I've seen that appears relevant:
* Fix an incompatibility issue between GPFS 3.5 and GPFS 3.4. When both versions coexist, nodes running GPFS 3.4 can experience assert like "logAssertFailed: this == newDesc sgdescio.C".
* Fix an issue in a mixed version cluster, where a node running running GPFS 3.4 or older failing in a small window during mount could cause spurious log recovery errors.
* Fix assert "MSGTYPE == 34" that occurs in pre and post-126.96.36.199 mixed multicluster environment.
* Fixed the backward compatibility error in reading data across node on different versions. This is needed if you are upgrading from 188.8.131.52 or lower version number to 184.108.40.206 or higher GPFS version.
I don't see anything that looks relevant in the 3.4 changelogs.
To me, this would seem to indicate that--of course barring undiscovered bugs--if I run clients with 220.127.116.11, it should be fine? Anyone have any experience with this?
Pinned topic mixed 18.104.22.168 and 22.214.171.124 clients with 126.96.36.199 servers?
Answered question This question has been answered.
Unanswered question This question has not been answered yet.
Updated on 2013-01-04T19:04:41Z at 2013-01-04T19:04:41Z by FelipeKnop
dmr1 270003YTCN12 Posts
Re: mixed 188.8.131.52 and 184.108.40.206 clients with 220.127.116.11 servers?2013-01-04T17:43:41ZThis is the accepted answer. This is the accepted answer.
- FelipeKnop 120000G51H