Cloned system 552 errors on aix 5.1 but not 5.2 on...
We have done extensive testing using EMC srdf to basically clone
rootvg's from running systems to our DR site. The aix 5.2 servers seem
to failover well but on aix 5.1 we get 552 boot errors.
EMC has a 552 error work around that calls for booting into maint mode
, importing the rootvg disk then rmdev the hdisks that have rootvg
defined. Seems with aix 5.1 the odm it pulls in keeps the original
servers hdisk locations which it can't find on the DR hardware, thus it
hangs. The removing of the definitions in maint mode causes it to re
discover the the devices and it boots.
Any ideas what as changed in aix 5.2 that could make this work ?
ApprenticeRating: 0%, 0 votes
Thanks for the information. How did you modify the ODM for aix 5.2 to
work? I understand the commands but we are worried if we modify the odm
on the source server it will cause a problem ?
We have worked with IBM and just here that bit level copies are not
supported . We are continuing to push IBM .
If you don't mind let me know if there is an easier way to contact and
we can compare notes .
Was this solution helpful? Show your Appreciation by rating it:
ApprenticeRating: 0%, 0 votes
privately with a working address and I'll re-send. Maybe I just can't
send messages to yahoo from gmail.
Thanks.
/cam
Was this solution helpful? Show your Appreciation by rating it:
ApprenticeRating: 0%, 0 votes
switches with no persistent FCIDs) and the story we got from IBM is
that there are fixes for this (dynamic tracking) in AIX 5.2 that aren't
in 5.1. We did the same thing you did to recover the systems initially,
but the problem has now been solved. IBM backported these fixes to 5.1
for us (and some other customers); they've been tested and they work.
You can change the scsi_id attribute in the ODM for the fscsi/dac/hdisk
devices and it will discover a new one and continue to boot. There are
some caveats to this, but it works well for the most part. We also
enabled persistent FCIDs in the switches which should have been done in
the first place. You're restoring through different switches so you
need the functionality in the fixes.
root@host:/home/cam# emgr -l
ID STATE LABEL INSTALL TIME ABSTRACT
===
1 S CFG_HA_51G 08/21/04 06:52:57 Dynamic Tracking and HACMP
fixes
2 S fastt_lib 08/21/04 06:53:58 FAStT libcfgscsi.a for dac
tracking
These were supposed to be released as APARs IY58854 and IY59851, but it
should be noted at at least one of these is a FAStT-specific fix, so
I'd ask IBM or search for an APAR for your particular hardware.
Let me know if you want any further information on the problem, I've
unwittingly become an expert in AIX and FAStTs.
/cam
Was this solution helpful? Show your Appreciation by rating it:
Can you Help with these Network Storage problems?
Tags | Blog | Repair Directory | Extended Warranty | Partners | Jobs |
