ML370 Windows NT 4.0 startup problem

[Aby]

I have a few Proliant ML 370 servers with SMART Array 431 controllers. 3x9.1 gb HDDs in them configured on RAID - 5.1 GB of physical memory. The Operating system is Windows NT4.0 (SP6a). 3 partitions. C: drive has a pagefile of 1.3 gb.

Whenever I start the server, during NT startup, it takes a long time to proceed from the Blue screen (30 mins to 1 hr).
Its happening with all the ML 370 servers (all of them have same config hardware and software wise).

Any suggestions or ideas ?

[Leonard Davison]

Aby...


Try adding the /sos switch to the appropriate line in boot.ini file.
This will cause the NT loader to list each device as it is starting
up.

If you find one particular device that is taking way too long to boot,
you might try changing it's startup (in control panel) to automatic.
Be careful, as this can render you system unbootable if you change
a critical device driver such as NTFS.

[Aby]

Hi Leonard,

Thanks for the early response.

I tried the /sos switch in boot.ini.

The server passes the ntdetect stage and reaches the blue screen where it shows the Operating system label, Service pack version and RAM size. It stays there for more that 30 mins. At this time can see a lot of activity on the disks, and the lights keep blinking alternately at fixed intervals on each disk.

[Leonard Davison]

Aby...

Is there anything else connected to the SCSI busses on the machines? (Tape Drive, Syquest etc.)

Is there any information in the System or Application Eventlog showing Timeout on SCSI devices?

I know that NT with large amounts of RAM take a while to initialize but 30min is a long time...How about Driver and BIOS versions for the Array controller...Updated to the latest?..

[SAKET_5]

Hi Aby,

A couple of things that I would be curious to check:

1. Perform a server hardware diagnostic. I believe you can use HP server diagnostic tools by downloading it from their web site.

2. Check if there are memory/parity errors using the same set of tools as above?

3. Check to make sure that the FW on SCSI drives are up to date.

4. Check to make sure that you are running the latest device drivers? A bug in a specific device driver may cause a long delay in loading it up and hence other dependent drivers/services load up will be delayed consequently.

5. Are there any applications that you are running on the server that are set to start up automatically that might be taking a long time to start. Take a look at the application log on NT service initialilsation failed or request timeout errors/warnings!

6. Any odd System event log messages?

7. Can you try moving your pagefile to a less busy disk and see if this speeds up your boot up?

Hope, this helps and dont forget to assign points:)

Let us know how you went,

Good luck!

[Aby]

The firmware has been updated to the latest available on HP website.

There are no errors about SCSI timeouts on device driver initialisation errors in NT event logs.

As of now, all the 13 servers we have are affected.

We solve the problem temporarily by wiping the C: drive and reinstalling the OS.

After a fresh install it works fine for a week or so.

After that period when it is restarted, it stays on that blue screen for a long time.

The only application we are running on the server is Lotus Domino r5.The C: has the windows nt 4.0 OS only.

[Leonard Davison]

Aby...

Hmmm...I am assuming they are part of an NT Domain....

What happens if you unplug the network cable during boot? Does it speed up?

Are you running adapter teaming on the NIC?

What happens if you pull out 4GB of RAM, does the speed increase?

This may be completely unrelated to the HD's and controller.

Len

[Roger Faucher]

Hi:

Are you running NT4 Enterprise Edition? Otherwise, I believe the maximum memory is 4GB. Have you tried lowering one of them to 4GB RAM? Try raising the pagefile up to the maximum (4095MB).

HTH

Make a great day!

Roger