RE: New SAN Selection Question

[gsweers]

Ease of Configuration: EQ 30 mins mapping drives if you have never
played with it before

 

Total functionality: snapshot, SAN, NAS, replication, etc.  Look at
costs for each of these.  Only a few do it all out of the box without
addt'l costs.

Evaluation of compliance, security, encryption needs

 

Integration and Support:  VMWARE, MS Virtual, etc native.

 

Fault Tolerance and Recoverability:  Equalogic was pretty dang
impressive when they pulled all the drives out, swapped them around, and
then pulled both redundant powers for a hard shutdown and it powered up
and everything was instantly available.  Never seen this with anyone
else.

 

Speed and Throughput:  SAS or Fiber /not necessarily faster.   EQ gets
faster as you add chassis since it forms a matrix between units.  Its
not just adding storage, its scaling everything..

 

Scalability: Cost, Performance, Simplicity

 

Replication:  EVALUTATE total storage requirements.  Differences between
entire disk rep and volume rep.  Many require between 20% and 150% of
the disk space to replication reserved.  So 1 TB repl may actually
require 2.5TB to do that.  

 

Sorry EQ fan here, used Netapp and not impressed with mgmt interface or
feature list when I compare it up to EQ.  I have 2 friends who use EMC
and both have said that they are good units, but require a degree to
fully manage it and it does not scale well enough for them without
significant cost each time.

 

One thing about EQ, with each chassis you can only do 1 Raid Type across
all of the disks.  Not a problem with the perf it puts out in most
cases, but something to be mindful of.

 

Greg

 

 

From: Jeffrey Showen [mailto:[EMAIL PROTECTED] 
Sent: Sunday, June 29, 2008 3:58 PM
To: NT System Admin Issues
Subject: New SAN Selection Question

 

I'm looking for a methodology to evaluate new Storage Area Network
solutions for our enterprise environment.  We have tight power, weight,
and support requirements and have narrowed the field to 8 vendors based
on a variety of these (and other) threshold requirements.  The challenge
now is that the vendor submissions/solutions are mostly so close to each
other that a paper evaluation fails to significantly differentiate them.
We are on a tight schedule (management - don't ask) and it looks like we
will end up testing I/O on solutions from all 8 vendors (Dell, EMC, IBM,
NetApp, LeftHand Networks, Sun, Overland, HP).

 

We are an ESX shop - is it enough to just configure the eval SANs (the
same) for our environment and then run Iometer tests from a client to
the primary SAN(s) or is there a better or more thorough approach?  We
want to use mirroring/replication at the SAN level between redundant
SANs for high availability but that can get expensive so I am open to
any ideas here as well.

 

I know this is a big question with sketchy details but there are a lot
of moving parts and I was just wondering if someone had done this and
could provide insight.

 

Thanks

 
Jeff Showen 
iGov Systems Engineer
Tampa, FL 33619

 

 

~ Upgrade to Next Generation Antispam/Antivirus with Ninja!~
~   ~

RE: New SAN Selection Question

[Martin Blackstone]

Be sure to take a close look at the applications that are available for each
appliance.

Can the appliance do both SAN and NAS?

What are their provisioning tools like? How easy can you provision a LUN?

Backup and Recovery. Compliance?

What are their Exchange / SQL tools like? Or whatever other applications you
may use.

Watch their snapshot technology and keep a close eye on performance impact
during a snapshot.

 

With ESX, we were originally running over ISCSI but I've recently moved
everything to NFS storage. I find it to be much more flexible.

VMware is also a prime candidate for deduplication, so ask if your vendor
offers dedupe technologies.

 

 

 

 

From: Jeffrey Showen [mailto:[EMAIL PROTECTED] 
Sent: Sunday, June 29, 2008 3:58 PM
To: NT System Admin Issues
Subject: New SAN Selection Question

 

I'm looking for a methodology to evaluate new Storage Area Network solutions
for our enterprise environment.  We have tight power, weight, and support
requirements and have narrowed the field to 8 vendors based on a variety of
these (and other) threshold requirements.  The challenge now is that the
vendor submissions/solutions are mostly so close to each other that a paper
evaluation fails to significantly differentiate them.  We are on a tight
schedule (management - don't ask) and it looks like we will end up testing
I/O on solutions from all 8 vendors (Dell, EMC, IBM, NetApp, LeftHand
Networks, Sun, Overland, HP).

 

We are an ESX shop - is it enough to just configure the eval SANs (the same)
for our environment and then run Iometer tests from a client to the primary
SAN(s) or is there a better or more thorough approach?  We want to use
mirroring/replication at the SAN level between redundant SANs for high
availability but that can get expensive so I am open to any ideas here as
well.

 

I know this is a big question with sketchy details but there are a lot of
moving parts and I was just wondering if someone had done this and could
provide insight.

 

Thanks

 
Jeff Showen 
iGov Systems Engineer
Tampa, FL 33619


~ Upgrade to Next Generation Antispam/Antivirus with Ninja!~
~   ~

Re: New SAN Selection Question

[Jeffrey Showen]

Yeah, I checked ISE out at SNW in Orlando as well - doesn't really fit
within our architecture but very impressive.  Besides, management wanted the
vendors on our list and no others - their view of "Tier 1" vendors.  I did
manage to get LeftHand and 3PAR on the list though I did get beat up over
it.

The ISE display was tight- 1 PB in 3 standard 72U racks and a starship blue
glow!  The self-healing drive packs are an interesting idea - I'll wait to
see how it works out as thay start collecting data on failure rates, etc.

LeftHand missed the submission cutoff and then bullied their way into our
storage survey anyway - very pushy indeed and frankly burning bridges before
they cross them.  We evaluated HP DL380s for a server project and they were
strong but didn't win and that is the LeftHand platform they are trying to
sell us (with SAN/iQ of course).  Dell and EMC both proposed entry level
SANs (AX4-5) - they are cheap and I can't seem to make management understand
you get what you pay for!

However, I don't need vendors, I need evaluation methodologies.  Any ideas
will be greatly appreciated.

Jeff

On Sun, Jun 29, 2008 at 8:22 PM, <[EMAIL PROTECTED]> wrote:

>
> I highly recommend adding Xiotech to the list. Their new "ISE" based SAN
> (specifically look at the Emprise 7000) is impressive. Played with it at
> Storage World in Orlando last month, and was highly impressed.
>
>
>
>
>
>   *"Jeffrey Showen" <[EMAIL PROTECTED]>*
>
> 06/29/2008 06:57 PM   Please respond to
> "NT System Admin Issues" 
>
>To
> "NT System Admin Issues"   cc
>   Subject
> New SAN Selection Question
>
>
>
>
> I'm looking for a methodology to evaluate new Storage Area Network
> solutions for our enterprise environment.  We have tight power, weight, and
> support requirements and have narrowed the field to 8 vendors based on a
> variety of these (and other) threshold requirements.  The challenge now is
> that the vendor submissions/solutions are mostly so close to each other that
> a paper evaluation fails to significantly differentiate them.  We are on a
> tight schedule (management - don't ask) and it looks like we will end up
> testing I/O on solutions from all 8 vendors (Dell, EMC, IBM, NetApp,
> LeftHand Networks, Sun, Overland, HP).
>
> We are an ESX shop - is it enough to just configure the eval SANs (the
> same) for our environment and then run Iometer tests from a client to the
> primary SAN(s) or is there a better or more thorough approach?  We want to
> use mirroring/replication at the SAN level between redundant SANs for high
> availability but that can get expensive so I am open to any ideas here as
> well.
>
> I know this is a big question with sketchy details but there are a lot of
> moving parts and I was just wondering if someone had done this and could
> provide insight.
>
> Thanks
>
> Jeff Showen
> iGov Systems Engineer
> Tampa, FL 33619
>
>

~ Upgrade to Next Generation Antispam/Antivirus with Ninja!~
~   ~

Re: New SAN Selection Question

[Mike Sullivan]

We recently purchased a new SAN and it came down to Dell EqualLogic and
Lefthand. Typically, Lefthand would win the price war but we were lucky that
Dell had just purchased EqualLogic and was closing out the PS100e's so they
beat Lefthand by a pretty good margin. Both have their pros and cons and it
came down to me just liking the pros of the EqualLogic over the pros of the
Lefthand. My biggest pro was the ease of use with the EqualLogic interface.
I have 200 machines across 5 offices and 2 field offices and only 1
assistant so I do not have time to be an expert. I also did not like how
pushy the Lefthand are reps were. To top it off when I told them the
EqualLogic won they basically said that I was making a big mistake and would
regret my decision.



On Sun, Jun 29, 2008 at 3:57 PM, Jeffrey Showen <[EMAIL PROTECTED]> wrote:

> I'm looking for a methodology to evaluate new Storage Area Network
> solutions for our enterprise environment.  We have tight power, weight, and
> support requirements and have narrowed the field to 8 vendors based on a
> variety of these (and other) threshold requirements.  The challenge now is
> that the vendor submissions/solutions are mostly so close to each other that
> a paper evaluation fails to significantly differentiate them.  We are on a
> tight schedule (management - don't ask) and it looks like we will end up
> testing I/O on solutions from all 8 vendors (Dell, EMC, IBM, NetApp,
> LeftHand Networks, Sun, Overland, HP).
>
> We are an ESX shop - is it enough to just configure the eval SANs (the
> same) for our environment and then run Iometer tests from a client to the
> primary SAN(s) or is there a better or more thorough approach?  We want to
> use mirroring/replication at the SAN level between redundant SANs for high
> availability but that can get expensive so I am open to any ideas here as
> well.
>
> I know this is a big question with sketchy details but there are a lot of
> moving parts and I was just wondering if someone had done this and could
> provide insight.
>
> Thanks
>
> Jeff Showen
> iGov Systems Engineer
> Tampa, FL 33619
>
>


-- 
Mike Sullivan
[EMAIL PROTECTED]

~ Upgrade to Next Generation Antispam/Antivirus with Ninja!~
~   ~

Re: New SAN Selection Question

[jeff . wilhelm]

I highly recommend adding Xiotech to the list. Their new "ISE" based SAN 
(specifically look at the Emprise 7000) is impressive. Played with it at 
Storage World in Orlando last month, and was highly impressed.






"Jeffrey Showen" <[EMAIL PROTECTED]> 
06/29/2008 06:57 PM
Please respond to
"NT System Admin Issues" 


To
"NT System Admin Issues" 
cc

Subject
New SAN Selection Question






I'm looking for a methodology to evaluate new Storage Area Network 
solutions for our enterprise environment.  We have tight power, weight, 
and support requirements and have narrowed the field to 8 vendors based on 
a variety of these (and other) threshold requirements.  The challenge now 
is that the vendor submissions/solutions are mostly so close to each other 
that a paper evaluation fails to significantly differentiate them.  We are 
on a tight schedule (management - don't ask) and it looks like we will end 
up testing I/O on solutions from all 8 vendors (Dell, EMC, IBM, NetApp, 
LeftHand Networks, Sun, Overland, HP).
 
We are an ESX shop - is it enough to just configure the eval SANs (the 
same) for our environment and then run Iometer tests from a client to the 
primary SAN(s) or is there a better or more thorough approach?  We want to 
use mirroring/replication at the SAN level between redundant SANs for high 
availability but that can get expensive so I am open to any ideas here as 
well.
 
I know this is a big question with sketchy details but there are a lot of 
moving parts and I was just wondering if someone had done this and could 
provide insight.
 
Thanks
 
Jeff Showen 
iGov Systems Engineer
Tampa, FL 33619

~ Upgrade to Next Generation Antispam/Antivirus with Ninja!~
~   ~

Re: New SAN Selection Question

[Kurt Buff]

We just purchsed (they're sitting in boxes on our server room floor,
right now!) two 4Tb Lefthand units. For us, it came down to ROI,
failover and integration with Windows/ESX. We'd also considered
Datacore, and they were close, but the numbers worked for Lefthand.
ROI was a bit more complicated than simply budget numbers, though. We
looked at ease of use, maintenance and a few other things, after
watching lots of demos, both online and f2f.

Needless to say, I'm stoked. Once we get these in place, we can make
the case for the Enterprise version of VMWare, with vmotion, etc.,
much more easily.

On Sun, Jun 29, 2008 at 3:57 PM, Jeffrey Showen <[EMAIL PROTECTED]> wrote:
> I'm looking for a methodology to evaluate new Storage Area Network solutions
> for our enterprise environment.  We have tight power, weight, and support
> requirements and have narrowed the field to 8 vendors based on a variety of
> these (and other) threshold requirements.  The challenge now is that the
> vendor submissions/solutions are mostly so close to each other that a paper
> evaluation fails to significantly differentiate them.  We are on a tight
> schedule (management - don't ask) and it looks like we will end up testing
> I/O on solutions from all 8 vendors (Dell, EMC, IBM, NetApp, LeftHand
> Networks, Sun, Overland, HP).
>
> We are an ESX shop - is it enough to just configure the eval SANs (the same)
> for our environment and then run Iometer tests from a client to the primary
> SAN(s) or is there a better or more thorough approach?  We want to use
> mirroring/replication at the SAN level between redundant SANs for high
> availability but that can get expensive so I am open to any ideas here as
> well.
>
> I know this is a big question with sketchy details but there are a lot of
> moving parts and I was just wondering if someone had done this and could
> provide insight.
>
> Thanks
>
> Jeff Showen
> iGov Systems Engineer
> Tampa, FL 33619

~ Upgrade to Next Generation Antispam/Antivirus with Ninja!~
~   ~

New SAN Selection Question

[Jeffrey Showen]

I'm looking for a methodology to evaluate new Storage Area Network solutions
for our enterprise environment.  We have tight power, weight, and support
requirements and have narrowed the field to 8 vendors based on a variety of
these (and other) threshold requirements.  The challenge now is that the
vendor submissions/solutions are mostly so close to each other that a paper
evaluation fails to significantly differentiate them.  We are on a tight
schedule (management - don't ask) and it looks like we will end up testing
I/O on solutions from all 8 vendors (Dell, EMC, IBM, NetApp, LeftHand
Networks, Sun, Overland, HP).

We are an ESX shop - is it enough to just configure the eval SANs (the same)
for our environment and then run Iometer tests from a client to the primary
SAN(s) or is there a better or more thorough approach?  We want to use
mirroring/replication at the SAN level between redundant SANs for high
availability but that can get expensive so I am open to any ideas here as
well.

I know this is a big question with sketchy details but there are a lot of
moving parts and I was just wondering if someone had done this and could
provide insight.

Thanks

Jeff Showen
iGov Systems Engineer
Tampa, FL 33619

~ Upgrade to Next Generation Antispam/Antivirus with Ninja!~
~   ~

Re: Hoping to find someone tonight...

[Kurt Buff]

Too late, unfortunately - I'm now at work dealing with downed air
conditioning, after putting the switches all back together again on
Friday night.

On Sat, Jun 28, 2008 at 6:13 PM, Andy Shook
<[EMAIL PROTECTED]> wrote:
> If not to late, would you please post the output of sh int  that is acting up> when your having the problem.  I've never seen a
> Cisco switch with all ports throwing a protocol error, just single ports
> here and there.  It "almost" sounds like a spanning-tree problem...or a
> couple of bad switches.  Call the TAC if you have smartnet or there less
> than a year old...
>
> Shook
>
> -Original Message-
> From: Kurt Buff [mailto:[EMAIL PROTECTED]
> Sent: Saturday, June 28, 2008 7:45 PM
> To: NT System Admin Issues
> Subject: Re: Hoping to find someone tonight...
>
> Sorry - what I wrote below is a little misleading.
>
> I had a console attached to the 2948T-48s - they were the ones
> reporting the protocol error. I didn't see much on the 3400cl.
>
> Kurt
>
> On Sat, Jun 28, 2008 at 4:43 PM, Kurt Buff <[EMAIL PROTECTED]> wrote:
>> The link light would appear for a short while, then disappear, and
>> output in the console window talked about a protocol error. For each
>> and every one of the ports to which a 2950T-48 was attached. And, for
>> one of those ports, I undid the cable for the 2950T-24, and hooked a
>> 2950T-48 into it. No go.
>>
>> Same exact cables, because the 2948G-L3 that we're replacing is less
>> than 1U away from the HP 3400cl that's replacing it. I just took the
>> cables from their ports on the 2948G-L3 and plugged them into the
>> 3400cl.
>>
>> I also power cycled all switches, in sequence. Nothing.
>>
>> Kurt
>>
>> On Sat, Jun 28, 2008 at 3:02 PM, Phil Brutsche <[EMAIL PROTECTED]>
> wrote:
>>> No link?
>>>
>>> No link at *all*?
>>>
>>> Even with known-good cables?
>>>
>>> Generally speaking I've only ever seen that when one or both sides of
>>> the link has bad ports.
>>>
>>> The only other time I've seen a Cisco switch not show an active link
>>> with a known-good cable is when the port is in the "shutdown" state,
> and
>>> I'm sure you've considered that possibility.
>>>
>>> Kurt Buff wrote:
 Their link lights aren't even lit. Nothing. Nada. I power cycled
> both
 the 2950Ts and the 3400cl, with no luck.
>>>
>>> --
>>>
>>> Phil Brutsche
>>> [EMAIL PROTECTED]
>>>
>>> ~ Upgrade to Next Generation Antispam/Antivirus with Ninja!~
>>> ~   ~
>>>
>>
>
> ~ Upgrade to Next Generation Antispam/Antivirus with Ninja!~
> ~   ~
>
> ~ Upgrade to Next Generation Antispam/Antivirus with Ninja!~
> ~   ~
>

~ Upgrade to Next Generation Antispam/Antivirus with Ninja!~
~   ~