Veritas Cluster Server and Storageworks connected to P660 running AIX 5.1

Veritas Cluster Server and Storageworks connected to P660 running AIX 5.1

Post by Stephen Wal » Thu, 06 Feb 2003 03:29:33

Has anyone ever tried the following

Cluster running Veritas Cluster  - Veritas Foundation Suite HA v3.4 on
a P660 running AIX V5.1, connected to a Brocade 2400 switch (firmware
2.1.9g) and then Compaq Storageworks HSG80 controller (MA8000 array,
ACS 8.7F)

The software is working OK and failing over except when the array is
physically disconnected - at this stage it hangs and recovers 10
minutes later.

I'm looking for any feedback at all - has anyone else ever tried this,
is there any other "gotchas" with the scenario above.

Thanks for any help at all.



Veritas Cluster Server and Storageworks connected to P660 running AIX 5.1

Post by Aamir Kh » Thu, 06 Feb 2003 18:03:57


I have done a lot of testing of VCS on AIX 5.1 on differing platforms,
whilst testing there was some fun involved. But generally all worked.

There are some fixes on VCS what version are you running?

As for the hang? what do you get in the error report, I have been
testing using SSA arrays and later EMC Syms. With HACMP and Powerpath
It seems to work, there was a hang initially but that was due to the
HBA and powerpath s/w not licensed correctly.

I am still testing VCS at the moment setup being H70's and SSA D40
drawers. Lots of NICs, more than HA, based upon the "heartbeat" being
so different thatn HACMP.

I will dig out some other results from the test, I am just glad
someone out there is doing the same, although an avid HACMP fan, VCS
has definately "on the latest fixes and version" impressed a sceptic

Did some tests on foundation suite but just regarded it too expensive
as LVM is free.

quirks to follow when I dig them out.

kind regards Aamir


1. AIX 5.1 connected to StorageWorks EMA12000 SAN

Hi All,

I've recently purchased a new IBM p650 to replace an older 6M1. The
6M1 was connected to our StorageWorks SAN via a pair of Compaq/HP AIX
PCI FC adapters (197819-B21) running Securepath 2.0A all on AIX 4.3.3.
(The FC HBAs are actually Cambex FC1000's)

To migrate to the p650, which is running AIX 5.1 ML03, I varied off
the volume group, exported the vg, shutdown the 6M1, migrated the
cards into the new server, installed the SecurePath 2.0B software, and
did a cfgmgr. Everything appeared to be ok. I saw my 3 hdisks,
(hdisk2, 3, & 4), and after an importvg all my logical volumes came
back, I mounted them, and everything looked good.

Now, I'm occassionaly getting errors in errpt such as (I/O ERROR
DETECTED BY LVM), that coincide with errors in our applications error
log stating a failure to do an io (reads mostly). In general, I know a
lot of that type of error means a potential disk/hw problem, but we
only see one every couple days. However, it does cause problems for
the application (Sybase) so I need to get this fixed. fsck of the
affected filesystem doesn't turn anything interesting up.

Is anyone else doing something similar from a config standpoint? Is my
migration procedure potentially at fault? Should I have done anything
else instead? Any input, help much appreciated. IBM and HP/Compaq are
having a tough time deciding whose problem this may be, so I'm getting
desperate to find some answers.


2. What's FUD

3. AIX 5.1.B and 5.1.C and AIX 5.1.D

4. Recommended Graphics Adapters

5. Sun Cluster 3.0 vs Veritas Cluster Server

6. What is VM worth?

7. VERITAS cluster server vs. SUN cluster -- Any recommendations ?

8. Does It Matter Which Card is eth0 and which eth1?

9. RED HAT Advanced Server vs VERITAS Cluster Server

10. AIX 4.3.3 to AIX 5.1 with RAN

11. Veritas Cluster versus SUN Cluster