| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 2062.1 | Hints | NNTPD::"[email protected]" | Dave Cherkus | Thu May 15 1997 15:00 | 32 |
| >
> Hi,
>
> I'm doing a rolling upgrade on a Turbolaser cluster, and the member that I
> have upgraded
> panics on reboot.
>
> The message I get is:
>
> panic(0); memory channel - timed out waiting for response from node.
>
> Does this point to a broken mc card?
Yes, or some other form of unhappiness at the memory channel layer.
Strange, since presumably you didn't change the cables, jumpers
or position of the board.
> Has anyone got Production Server running on TCR1.4+DU4.0B?
I believe this was qualified in our lab, but I'm no authority.
> Also, the release notes mention putting rm-driver-debug-error=1 in the
> sysconfigtab. Which stanza is this, because it isn't recognised in rm.
"rm_spur".
> The remaining member is still at TCR1+DU3.2g awaiting it's partners return.
> Regards,
> Richard.
>
[Posted by WWW Notes gateway]
|
| 2062.2 | Curiouser..and curiouser..(and more curious) | NNTPD::"[email protected]" | Richard Carter | Thu May 15 1997 16:30 | 8 |
| If I shutdown the 3.2g machine, the v4.0b machine will boot.
Is there a conflict between TCRv1 and TCR1.4?
I have not run dlm_enable, which I believe may cause such a conflict.
Regards,
Richard.
[Posted by WWW Notes gateway]
|
| 2062.3 | MC the same rev . | NNTPD::"[email protected]" | Richard Carter | Thu May 15 1997 16:38 | 15 |
| Hi Dave,
Thanks for your note.
I wrote -1 before I saw your reply.
MCS have very kindly replaced both MC cards and the cable to no effect.
So, this may point to a software problem.
Who should I ask next?
(I will log this with my CSC tomorrow; but any help would be nice)
Regards,
Richard.
[Posted by WWW Notes gateway]
|
| 2062.4 | worked oK for me | KYOSS1::GREEN | | Fri May 16 1997 09:58 | 6 |
| I recently did "SIMULTANEOUS" cluster upgrade from:
3.2f, TCR 1.0--> 3.2g, TCR 1.0--> 4.0a, TCR 1.4--> 4.0b, TCR 1.4
All went well. We followed the documentation and had NO problems.
The config was 2 2100s in a virtual hub.
dick
|
| 2062.5 | 2100 O.K. | NNTPD::"[email protected]" | Richard Carter | Mon May 19 1997 06:42 | 11 |
| Hi Dick,
Thanks for the info.
I have also upgraded 2100's without a problem, by using rolling upgrade.
Regards,
Richard.
[Posted by WWW Notes gateway]
|
| 2062.6 | MC rev 11 or 14? | NNTPD::"[email protected]" | Richard Carter | Mon May 19 1997 13:15 | 26 |
| Hi,
I was working on a different TruCluster/TurboLaser customer today.
This one is up and running O.K.
There seems to be two differences between the working cluster and
the one that panics:
1. Memory channel card Rev.
The panicing systems have Rev 11
The O.K. systems have Rev 14.
2. The O.K. system was a fresh install not an upgrade.
What is the following sysconfigtab for and when is it enabled. i.e.
during an
upgrade or afterwards?
dlm:
dlm_disable_grptx = 0
The customer is currently deciding whether to abandon MC altogether,
and go ASE only. Is there any debugging info that could help?
Regards,
Richard.
[Posted by WWW Notes gateway]
|