format taking too long

[admin1979]

Hello,

We have a TRU64 TRU Cluster with disk enclosures attached.
Recently we added a new disk in slot to be used as a SPARESET and did the following so far.

1. run config

2. run hsutil

- Select 0 to format
- say yes to start the formatting.

Normally the formatting takes couple of hourse to complete.
But this time, it is going on since over 24 Hrs now. And I feel this is no normal. Correct me if I am wrong.
I could still see the LED blink indicating the formatting is going on.
Also the HSUTIL is just displaying the message of (formatting) activity started on date-time.

I still can not get the HSUTIL command prompt back as the formatting is ON.
Not much info I could find for HSUTIL format option.

Now the question is how do I terminate this process or should I simply wait for it to complete?

Plz suggest.

admin

[Rob Leadbeater]

Hi admin,

What storage unit are you using ?

Cheers,

Rob

[Pieter 't Hart]

looking at the utility "hsutil" I assume you are using a HSZ storage controller?
You used the config utility to add the disk to the storage configuration as member of the spareset?

There was no need to run format (low level format) at all.
If preferred, you should only have used the "initialize disk" to create metadata information on the disk.
But for a spareset member I don't even think that would have been necessary.

That formatting takes so long may indicate the disk is faulty?

[admin1979]

Thanks for the replies.

Rob : It is HSZ Storage Controller.

Peter : This is completely new thing to me to know that there was no need to run low level format at all.
I suppose every new/supposed to be new disk which is to be reused as a new , should be formatted.

In fact for every new disk this is what we follow,

1. Add disk
2. run config
3. run hsutil
Select Format option
4. initialize disk_number
5. add spareset disk_number

Is it not required? Or it is only required if we are adding it to specific storage set other than spareset?

Please clarify.



Thanx,
admin

[Rob Leadbeater]

Hi admin,

It's been a while since I've used a HSZ based system, but I don't recall ever having to use the format command...

Cheers,

Rob

[Pieter 't Hart]

If you are adding replacing HP disks, there is no need to run format.

You can use "run config" to add the disk to the controllers' configuration or by CLI command like
> delete DISK31500
to delete the reference to the old disk
physically replace the disk
> add disk DISK31500 3 15 0
> initialize DISK31500
where 3 15 0 is Port Target Lun
> add spareset DISK31500

[admin1979]

Thanx Rob n Pieter!!

Next time onwards I will be using only run config (as it does the job of add disk command for me) and then simply initialize.

Thats great!!

But the problem is we cannot use command line interface. Everytime we have to get a console access to HSZ.
Previously it used to be like, run following command and get HSZ70 command prompt.

/usr/lbin/hsxterm5 -b 0 -t 1 -l 5

But few months back, when I tried to login to it, I ended up in FMU> command prompt rather than HSZ>.
Reason still unknown to the date.
Hence now we must connect console cable to get HSZ70 for any maintenance work :-(

Any idea?


Thanx,
admin

[Pieter 't Hart]

FMU is the field management utility of the HSZ.
You would have be the same response if you entered the "run FMU" command from the console cable.

type "exit" to exit FMU.

[admin1979]

No , unfortunately EXIT command is not working here.
EXIT simply comes out of the whole utility to get the unix command prompt.
Thats really strange!!

[Pieter 't Hart]

did you really used exit with capitals or not?

hsxterm may intercept the "exit" command.
but as it is unix it will be case sensitive.
try "EXIT".

just for the record
fmu is not Field but Fault Management Utility

[admin1979]

Ohh I tried both ....many times....but no luck..... :( it somehow is stucked up in that interface.

[Pieter 't Hart]

do you have a redundant controller?
then you can use "restart this_controller" from the console port.
when "this" controler is restarted restart the other "restart other_controller".

[admin1979]

Yes its a dual controller setup ,

root@pluto> clu_get_info
Cluster information for cluster host123

Number of members configured in this cluster = 2
memberid for this member = 1
Quorum disk = dsk12h
Quorum disk votes = 1

Information on each cluster member

Cluster memberid = 1
Hostname = pluto.com
Cluster interconnect IP name = pluto-ics0
Member state = UP

Cluster memberid = 2
Hostname = mars.com
Cluster interconnect IP name = mars-ics0
Member state = UP


Its a cluster in-use right now. would it not affect the users if we simply restart the controllers?
Is there any better way?

Thanx,
admin

[Pieter 't Hart]

if correctly configured, redundant controllers should take over i/o, so there is little impact on users.

other option may be in last post (Nick) from this thread:
http://forums13.itrc.hp.com/service/forums/questionanswer.do?admit=109447627+1283330603800+28353475&threadId=1390139

so try
/usr/lbin/hsxterm5 -b 0 -t 1 -l 5 "exit"

[Pieter 't Hart]

strange ? this post is also from admin1979?

[admin1979]

Like I stated before, this prob is there since long time.
But I somehow missed what Nick suggested.
This time I would give it a try to it.
Will update soon, thanx.

[Pieter 't Hart]

I found another post from you regarding the same problem.
And understand it's there for a long time.

If the command-line option does not help you'll have no option but to retart the crontrollers on a t a time.
do "show this_controller" to verify the
-------------
Configured for MULTIBUS_FAILOVER with
In dual-redundant configuration
-----------------
If you by chance can locate the controller that handles the hsxterm session, you may only need to restart this one.
It is the current active controler for "b 0 -t 1 -l 5"

[Pieter 't Hart · Accepted solution]

I found this
http://forums11.itrc.hp.com/service/forums/questionanswer.do?threadId=440328

MarkS said: To exit I proceeded exit with v. Worked for me.

There are some other suggestions From Uwe.

[admin1979]

Heyy Pieter , that thread really helped me out.

CTRL thing did work and the problem got solved. The HSZ70 prompt is back.
Its really wonderful that you could find a similar post with same problem.
I remember I too tried to search it sometime back but could not find related post.
But somehow you could find it.
Anyws thanx a ton for your help. You been always helpful. Hope u have got few points from me already. :-)

Thanx,
admin

[Pieter 't Hart]

thanks for the points.

your quick response to suggestions also helps a lot in this process.

regards,
Pieter

[Pieter 't Hart]

don't forget to close the other threads.

hope you get the format issue solved.
maybe the running fru got in the way?

[admin1979]

Yes.
After nearly 24+ hrs, the format process came out with a coninuous beep. And the disk was showing amber LED.
So it was clear that the disk was faulty.
As I have already ran delete disk , I simply took out the disk out of the array and replaced with new one.
The formatting of this new disk completed in 45 minutes.
So thats it.
Thanx Pieter n all once again and yes I would close d old open cases soon.

Regards,
admin

[admin1979]

The old disk is removed n replaced with new one as mentioned.