SUMMARY: clu_upgrade won't finish

From: Rick Beebe <richard.beebe_at_yale.edu>
Date: Wed, 10 Oct 2001 11:56:18 -0400

That was a fast one. Doreen's answer, below, gave me the right hint. I
was blithely editing /.rhosts but we'd changed root's home directory to
/root. Moving /.rhosts into /root/.rhosts fixed it. Now the switch
command says it finished. I guess it was unable to do the switch on the
other node. Since this is my first 5x patch, I didn't know what it was
supposed to do so I didn't know what I was missing! There are probably a
bunch of other commands that wouldn't work right either.

Thanks Doreen.

--Rick

-------- Original Message --------
From: "Alongi, Doreen" <Doreen.Alongi_at_compaq.com>
To: Rick Beebe <richard.beebe_at_yale.edu>

Hi Rick,

We had the exact same problem here. I don't recall
exactly what was wrong but it was something with our
/.rhosts file. It had been changed from when the cluster
was initially created and something was messed up.

Good Luck,
Doreen

-----Original Message-----
From: Rick Beebe [mailto:richard.beebe_at_yale.edu]
To: Alpha OSF Managers

Yesterday I tried installing patch kit 3 on a two-node cluster running
Tru64 5.1.

Everything went fine until, near the end, when you run "clu_upgrade
switch". The output from that was:

-------------------------
This is the cluster upgrade program.
You have indicated that you want to perform the 'switch' stage of the
upgrade.

Do you want to continue to upgrade the cluster? [yes]:
Initiating version switch on cluster members
.Marking stage 'switch' as 'started'.

kill: 531895: no such process
-------------------------

I didn't know if the kill error was normal or not, so I rebooted both
members as instructed. They both _seem_ to be running on the right
kernel now. Their sysconfigtabs look okay as best I can tell. But the
switch stage has not been marked completed and 'clu_upgrade clean'
refuses to run.

Anyone have any ideas? After perusing clu_upgrade (which is, to me, a
very obtuse KSH script) it looks like the kill error is caused by a call
to "clu_release_lock." I can't figure out why and I actually haven't
been able to figure out what the script is doing between the
clu_get_lock and clu_release_lock.


-- 
  
_______________________________________________________________________
    Rick Beebe                                            (203) 785-6416
    Manager, Systems & Network Engineering           FAX: (203) 785-3481
    ITS-Med Production Services                   Richard.Beebe_at_yale.edu
    Yale University School of Medicine
    Suite 214, 100 Church Street South, New Haven, CT 06519
  
_______________________________________________________________________
Received on Wed Oct 10 2001 - 15:51:16 NZDT

This archive was generated by hypermail 2.4.0 : Wed Nov 08 2023 - 11:53:42 NZDT