TruCluster
cancel
Showing results for 
Search instead for 
Did you mean: 

tru64 5.1b cluster not booting after installation

dominic_7
Advisor

tru64 5.1b cluster not booting after installation

Dear all ,

I created a tru64 5.1b cluster using a single GS60E alphaserver tru64 5.1B ,PK3 .
The server normally boots from SAN disks on a
HSG80 ACS 8.7 0P controller .

I added a new 36 GB ,disk , pertitioned it
into seven disks at the HSG level .
Used the various resulting partitions for
Cluster-root, cluster-usr , cluster-var ,
quorum ,member1 boot etc .

I was able to use clu_create to create the cluster sucessfully .
However after the installation when the system rebooted it gives the error as follows:

DRD cancelling register against 109 due to expired timer retrying operation .

CAM_logger scsi event packet

cam_logger : hardware_id= 109 bus2 target 1 lun145.
cdisk_handle_pr_ccb
request cancelled due to errors
Hard error detected
hardware id= 109
command timed out .


The unit that i had assigned on the HSG80 storage is named D145 , with ID =145 ,
I have checked the zoning on the SAN switch and it seems OK .
Permissions to the units are all OK .
connection type is tru64 .

Please suggest .

10 REPLIES
Uwe Zessin
Honored Contributor

Re: tru64 5.1b cluster not booting after installation

ACS V8.7-0 is very old. I don't know what data you put on D145, but I have seen problems, too, after initial setup of the cluster.

I suggest you first upgrade the patch level of your HSG environment:
http://h18000.www1.hp.com/products/storageworks/softwaredrivers/acs/index.html

Note that you have firmware cards with DRM functionality in them, so your environment might look a bit more complex. Make very sure you know what you are doing...
.
dominic_7
Advisor

Re: tru64 5.1b cluster not booting after installation

I dont think it is a HSG80 fw issue , snce i have another GS60E booting out of the same storage, using a partitioned HDD .

regards
Uwe Zessin
Honored Contributor

Re: tru64 5.1b cluster not booting after installation

Here is a comment from patch V8.7-2:

"On Tru64 Hosts (which supports the SCSI-3's implementation of Persistent Reservation) may experience an occasional LUN hang when a storage unit is being added utilizing a partition of a storageset."
.
Ralf Puchner
Honored Contributor

Re: tru64 5.1b cluster not booting after installation

Pleae check:

hardware_id= 109 bus2 target 1 lun145

If controlling the requirements (Firmware), cabling and hwmgr information you will get the information what is going wrong. It seems pure hardware
Help() { FirstReadManual(urgently); Go_to_it;; }
dominic_7
Advisor

Re: tru64 5.1b cluster not booting after installation

Thank you , well lun 145 , hardware id109 ,is the very unit that i have assigned as Member 1 boot disk .
Rhe cluster got formed using this unit as Member 1 boot, now while booting i boot from this very same hdd ,and it gets stuck at "starting CMS" with the SCSI timeout error

regards

Dominic


Ralf Puchner
Honored Contributor

Re: tru64 5.1b cluster not booting after installation

I've checked with similar messages and it indicates there is a firmware/cabling/setup problem. Most of the problems are EMC related.

So please check hardware and recommendation (firmware) etc.

It seems we can not solve the problem by talking about it ;-)
Help() { FirstReadManual(urgently); Go_to_it;; }
dominic_7
Advisor

Re: tru64 5.1b cluster not booting after installation

Thanks to all who replied .

As i had mentioned , i had selected the LAN interface (tu0)as the cluster interconnect .

There was an unused memory channel card in the system .
Today i removed the memory channel card ,and
the cluster came up .

Can anyone explain why ?

thanks

Dominic
Ralf Puchner
Honored Contributor

Re: tru64 5.1b cluster not booting after installation

if the members can not communicate via interconnect the hardware databases will not be synced leading to access problems.

Maybe cluster was configured to use the memory channel prior to the tulip cards. But this is only a guess without an analyze of the sys_check output.

If you are really interested in open a call within the support center for diagnose.
Help() { FirstReadManual(urgently); Go_to_it;; }
dominic_7
Advisor

Re: tru64 5.1b cluster not booting after installation


This is a single member cluster , so its just
a single member , during install i selected
LAn Interconnect , bcos if i selected "none"
for cluster interconnect , it would just exit and come to the unix prompt ,(some bug with Tru64 cluster ?).


Well regarding the cluster previously using Mem channel - i took a fresh 9GB HDD and did a
fresh install of tru64 5.1B with PK3 .
I then repartitioned the HSG based disks which
were being used for CFS etc .
I still faced the same problem .
I was selecting the LAN interconnect while installing the cluster .

After i removed the Memory channel card , the
cluster just booted normally.

So the memory channel though it was unused
was causing some issue .
What what pzzles me is that i have another
GS60 E in cluster with a ES80 running out
of the same storage , and this GS60 till recently also had an unused Memory channel card, but an active LAN cluster interconnect
but it did not show any such problems .

sincere regards

dominic Fernandes
Ralf Puchner
Honored Contributor

Re: tru64 5.1b cluster not booting after installation

A TruCluster always requires an cluster interconnect, so "none" will end setup that is correct (maybe time to study the cluster install guide).

It will be wise to be sure the HSG80 was visible and the discs properly configured (hwmgr) prior to creating the cluster because creation of the cluster depend on the information within the hardware databases. This includes checking the firmware revisions of each component.





Help() { FirstReadManual(urgently); Go_to_it;; }