IBM X336 server with RAID1 configuration formats all HDDs on newhardware installation

[Sid B]

(apologies for the cross-post)

Hi,

I am new to my company and have been asked to recommend a new
redundancy / backup plan for our online server. I've talked to the
other folks in the company and they have been describing an usual
feature of the server wherein if we try to install ANY new hardware
like a new harddisk, the system automatically formats everything.
(apparently this "resetting feature" kicks in even if we add more RAM
pieces!)

Now, I'm not too sure what portions of this are correct and what are
not. As this is our LIVE online server and we are serving hundreds of
thousands of documents on it at any point, therefore I am hesistant to
"try anything" before fully understanding all aspects.

The specifications are as follows:
CPU: Intel(r) Xeon (tm) cpu 3ghz (Dual processor)
Speed: 3 GHz
RAM: 4 GB
HDD: 2 x 140 GB (on RAID1)
Type: ADAPTEC RAID 1 SCSI Disk Drive
Server contains: Online Application and Database (SQL Server 2005)
Partitions: C: 30 GB, E: 110 GB
SCSI controller: (not known at this point)
The problems we're facing are as follows:
(a) Because of the low partition in C drive, we regularly have to run
maintenance tasks on the server which if we dont leads to our site
being inaccessible to our users.
(b) Our database is stored on E drive (about 60 GB of data). There is
a performance issue which our users face, but we are unsure of what we
should do to solve the problem.

I'm looking to understand:
(1) the automatic hard-disk formatting issue and how to solve it
(2) how we can upgrade our hardware in the most optimum fashion

Thanks!
Sid :)

[nik Simpson]

Sid B wrote:

(apologies for the cross-post)

You'd be better off asking on comp.arch.storage, or better yet IBM
support, because it definitely shouldn't be doing this.
--
Nik Simpson

[Del Cecchi]

"nik Simpson" <ni***@knology.netwrote in message
news:a9***************************@KNOLOGY.NET...

Sid B wrote:

>(apologies for the cross-post)

You'd be better off asking on comp.arch.storage, or better yet IBM
support, because it definitely shouldn't be doing this.
--
Nik Simpson

Since he is apparently running MS software, shouldn't he ask them? Or
is that support provided through IBM?

[nik Simpson]

Del Cecchi wrote:

"nik Simpson" <ni***@knology.netwrote in message
news:a9***************************@KNOLOGY.NET...

>Sid B wrote:

>>(apologies for the cross-post)

You'd be better off asking on comp.arch.storage, or better yet IBM
support, because it definitely shouldn't be doing this.

Since he is apparently running MS software, shouldn't he ask them? Or
is that support provided through IBM?

Can't be the OS, since when the reformatting takes place, the OS isn't
even loaded, seems more like a misconfigured RAID BIOS or hardware
problem, so I'd start with IBM support.

--
Nik Simpson

[Erland Sommarskog]

Sid B (si*******@gmail.com) writes:

I am new to my company and have been asked to recommend a new
redundancy / backup plan for our online server. I've talked to the
other folks in the company and they have been describing an usual
feature of the server wherein if we try to install ANY new hardware
like a new harddisk, the system automatically formats everything.
(apparently this "resetting feature" kicks in even if we add more RAM
pieces!)

That sounds scary! Well, my expertise is SQL Server, so I cannot comment
on that part.

The specifications are as follows:
CPU: Intel(r) Xeon (tm) cpu 3ghz (Dual processor)
Speed: 3 GHz
RAM: 4 GB
HDD: 2 x 140 GB (on RAID1)
Type: ADAPTEC RAID 1 SCSI Disk Drive
Server contains: Online Application and Database (SQL Server 2005)
Partitions: C: 30 GB, E: 110 GB
SCSI controller: (not known at this point)
The problems we're facing are as follows:
(a) Because of the low partition in C drive, we regularly have to run
maintenance tasks on the server which if we dont leads to our site
being inaccessible to our users.
(b) Our database is stored on E drive (about 60 GB of data). There is
a performance issue which our users face, but we are unsure of what we
should do to solve the problem.

The hardware configuration is not that impressive, but that does not
mean that it's insufficient. Your problem could just as well be a few
misbehaving queries, and fixing those, or just adding some indexes
could make your users very happy.

There are several ways to examine this. One way is to speak with users
to see if there are any specific operations that are slow. Another is
to ser up a trace that captures statements that runs for, say, more
than 100 ms, and the analyse that data.

There is a point when buying more hardware is cheaping than tuning,
particularly if tuning means lots of code to rewrite. But there is
also the other end when better hardware to some extent masks as
poorly implemented application, which you can get a lot better performance
from with small efforts.

--
Erland Sommarskog, SQL Server MVP, es****@sommarskog.se

Links for SQL Server Books Online:
SQL 2008: http://msdn.microsoft.com/en-us/sqlserver/cc514207.aspx
SQL 2005: http://msdn.microsoft.com/en-us/sqlserver/bb895970.aspx
SQL 2000: http://www.microsoft.com/sql/prodinf...ons/books.mspx

[Terje Mathisen]

Erland Sommarskog wrote:

Sid B (si*******@gmail.com) writes:

>I am new to my company and have been asked to recommend a new
redundancy / backup plan for our online server. I've talked to the
other folks in the company and they have been describing an usual
feature of the server wherein if we try to install ANY new hardware
like a new harddisk, the system automatically formats everything.
(apparently this "resetting feature" kicks in even if we add more RAM
pieces!)

That sounds scary! Well, my expertise is SQL Server, so I cannot comment
on that part.

[snip]

There is a point when buying more hardware is cheaping than tuning,
particularly if tuning means lots of code to rewrite.

In this particular case the "feature" is a horrible bug, and you should
run (not walk! :-) to your nearest dealer of reasonable hardware and get
a new server immediately.

You then setup this new server appropriately, copy your application over
and replace the broken one.

The reason I'm calling it broken is simply this:

With such a mis-feature, any kind of hardware error, like a RAM dimm
going bad and being mapped out, would cause the server to immediately
erase everything as soon as it had finished rebooting.
:-(

Terje
--
- <Te************@hda.hydro.com>
"almost all programming can be viewed as an exercise in caching"

[Del Cecchi]

"Terje Mathisen" <te************@hda.hydro.comwrote in message
news:Qv******************************@giganews.com ...

Erland Sommarskog wrote:

>Sid B (si*******@gmail.com) writes:

>>I am new to my company and have been asked to recommend a new
redundancy / backup plan for our online server. I've talked to the
other folks in the company and they have been describing an usual
feature of the server wherein if we try to install ANY new
hardware
like a new harddisk, the system automatically formats everything.
(apparently this "resetting feature" kicks in even if we add more
RAM
pieces!)

That sounds scary! Well, my expertise is SQL Server, so I cannot
comment
on that part.

[snip]

>There is a point when buying more hardware is cheaping than tuning,
particularly if tuning means lots of code to rewrite.

In this particular case the "feature" is a horrible bug, and you
should run (not walk! :-) to your nearest dealer of reasonable
hardware and get a new server immediately.

You then setup this new server appropriately, copy your application
over and replace the broken one.

The reason I'm calling it broken is simply this:

With such a mis-feature, any kind of hardware error, like a RAM dimm
going bad and being mapped out, would cause the server to
immediately erase everything as soon as it had finished rebooting.
:-(

Terje
--
- <Te************@hda.hydro.com>
"almost all programming can be viewed as an exercise in caching"

If it's a bug it wasn't there when it left the factory. Sounds like
some maroon screwed it up.

[Glen Herrmannsfeldt]

Terje Mathisen wrote:
(snip)

With such a mis-feature, any kind of hardware error, like a RAM dimm
going bad and being mapped out, would cause the server to immediately
erase everything as soon as it had finished rebooting.

I can see that as useful in some cryptographic applications.

-- glen

[Terje Mathisen]

Glen Herrmannsfeldt wrote:

Terje Mathisen wrote:
(snip)

>With such a mis-feature, any kind of hardware error, like a RAM dimm
going bad and being mapped out, would cause the server to immediately
erase everything as soon as it had finished rebooting.

I can see that as useful in some cryptographic applications.

As in macro-level quantum crypto?

If you try to inspect it, the bit(s) disappear!

Terje

--
- <Te************@hda.hydro.com>
"almost all programming can be viewed as an exercise in caching"