Solution Ideas

Teledude

Teledude

Teledude

Soldato
Joined
4 Dec 2003
Posts
2,847
Got a big presentation tomorrow morning for a new job in the city! :)

Anyway this is the brief just wondering if anyone has some useful ideas/suggestions :cool:

Client environment
1 site
150 servers (win2k3, 2k and linux)
550 users (some remote users)

Client requirements

1. To significantly reduce the current time to recover in case of a disaster down from 36 hours
2. A highly scalable, secure and resilient virtualised production environment
3. A long-term reduction in hardware, management and licensing costs
4. Improve ease of use of infrastructure management
5. Consolidation of Software
6. Increase server utilisation
7. Reduce server deployment time
8. Offer a more cost effective method to High Availability
9. Zero maintenance downtime
10. A highly effective and scalable DR/Business Continuity solution
11. Reduced Data Centre costs - Less power input and heat output
12. Reduce the costs and improve the ease of implementing future projects significantly (SQL migration and Exchange upgrade

Im thinking Bladecenters at Live and DR with some form of mirrored SAN infrastructure - possibly clustering somewhere. And a D2D based replicated backup solutuion. All based around ESX 3.5 and VirtualCenter. Am a bit lost as to points 3, 4 and 5 though :confused:
 
Live

6x c class HP Blades with dual quads and 32gig ram
4100 eva san with FC disks and a vls 6000 backup system

and replicated with doubletake for vi to two dl580 g5's at dr (quad quads and 40gb of ram in each - disk bays fully populated)

reasonable? :)
 
Teledude said:
Live

6x c class HP Blades with dual quads and 32gig ram
4100 eva san with FC disks and a vls 6000 backup system

and replicated with doubletake for vi to two dl580 g5's at dr (quad quads and 40gb of ram in each - disk bays fully populated)

reasonable? :)
Click to expand...

I don't think this much detail would be needed in a high level presentation. The solution should cover more generic technology terms such as SAN, Tape Library etc...

With regards to points 4 and 5:

4, Virtual Center
5, Citrix XenApp, XenDesktop
 
Teledude said:
Live

6x c class HP Blades with dual quads and 32gig ram
4100 eva san with FC disks and a vls 6000 backup system

and replicated with doubletake for vi to two dl580 g5's at dr (quad quads and 40gb of ram in each - disk bays fully populated)

reasonable? :)
Click to expand...

The EVA isn't brilliant for vmware, speak to Netapp would be my advice, they have some very nice integration with vmware, automatic replication of snapshots to a second SAN etc... (but Netapp snapshots have some issues with fragmentation so it's not perfect by any means).

We use the EVA when we need cheap, reliable SAN storage without too many features, but I wouldn't use it with vmware.
 
EVA and VMWare is fine. You can put in SAN replication using HP's Continuous Access but for that you need IP distance gateways (assuming replication between two sites?).

Wouldn't bother with Citrix as VMWare will tick most of you boxes (in fact all of them) just fill up a bladecente and migrate! You can use VMotion as well to move VM machiens on the fly meaning DR is easy!


M.
 
A brief, exciting!

Seems pretty much VM/Citrix/VDI-centric. I'd only have issues with 9... NDU is a bit of a hot topic :p

As a matter of interest, what sort of pay band does this job come in at? Apologies if it's too personal.

Ta
 
For points 3 & 4

Standardising hardware infrastructure like switches, routers, etc to one brand so that training/management is as straightforward as possible.

Also think about support contracts and how you could bring down the cost of those through using 1 supplier (like BT Lynx for example) that cover servers, printers, client PC's, switches etc
 
Shaz]sigh[;11895870 said:
A brief, exciting!

Seems pretty much VM/Citrix/VDI-centric. I'd only have issues with 9... NDU is a bit of a hot topic :p

As a matter of interest, what sort of pay band does this job come in at? Apologies if it's too personal.

Ta
Click to expand...

well i did get the job :D

and its upwards of 40k basic which im happy enough with :)
 
Congrats on getting the job!


1. To significantly reduce the current time to recover in case of a disaster down from 36 hours

A bit hard with one site because some disasters will wipe you out. Dual SAN at the same site may be an option. Consider iSCSI also.

2. A highly scalable, secure and resilient virtualised production environment

VMWare and a SAN.

3. A long-term reduction in hardware, management and licensing costs

Consolidation of servers via VMWare, consolidation of storage via SAN (less management), consolidation of licensing via vendor agreements and a use SAN software for mirroring rather than many different host based solutions

4. Improve ease of use of infrastructure management

If no SAN is in place today simply putting one in would improve overall management and backup

5. Consolidation of Software

As above, using a SAN to reduce replication technologies and utlise one piece of software for snaps, clones, etc..

6. Increase server utilisation

VMware & reduction of server count

7. Reduce server deployment time

VMWare & SAN. SAN for ease of provisioning.

8. Offer a more cost effective method to High Availability

SAN, SAN(s), Mirroring, VMWare path failover.

9. Zero maintenance downtime

VMotion and SAN to some degree, you may have to allow for some downtime.

10. A highly effective and scalable DR/Business Continuity solution

You're going to need two sites for this. A single SAN and tape retores is not effective.

EMC arrays, EMC Replication and VMware site recovery manager all work together today, other vendors do not have this.

11. Reduced Data Centre costs - Less power input and heat output

Vmware, SAN storage consolidation using high capacity drives for less important info

12. Reduce the costs and improve the ease of implementing future projects significantly (SQL migration and Exchange upgrade)

SAN provisioning, cloning, etc makes migration and projects relating to these two quite simple
 
Vanilla said:
10. A highly effective and scalable DR/Business Continuity solution

You're going to need two sites for this. A single SAN and tape retores is not effective.

EMC arrays, EMC Replication and VMware site recovery manager all work together today, other vendors do not have this.
Click to expand...

That's not the case, netapp ddi a demo of equivilent functionality to me a couple of months ago with snapmanager for virtual infrastructure.
 
Vanilla said:
Equivalent functionality to EMC Replication Manager or VMWare Site Recovery Manager?

.............as they're very different things
Click to expand...

Well VMware site recovery manager works with netapp, and the netapp replication functionality matched everything we saw from EMC. Disclaimer being I didn't pay much attention to the bits we didn't need, but we have a fairly complex environment and both netapp and EMC provided equivilent functionality for us.

I'm not in love with either product, both are flawed in their own ways (fragmentation of netapp snapshots being a delightful pain in the arse at the moment) but I'd say the functionality is equivilent.
 
bigredshark said:
Well VMware site recovery manager works with netapp, and the netapp replication functionality matched everything we saw from EMC. Disclaimer being I didn't pay much attention to the bits we didn't need, but we have a fairly complex environment and both netapp and EMC provided equivilent functionality for us.

I'm not in love with either product, both are flawed in their own ways (fragmentation of netapp snapshots being a delightful pain in the arse at the moment) but I'd say the functionality is equivilent.
Click to expand...

You know I hadn't even realised they announced support for it :(
 
bigredshark said:
Yeah, you can do it, but compared with the alternatives from other manufacturers it's not brilliant...
Click to expand...

I'd say it was easily comparable to any other SAN out there. Chuck some fibre channel disks in there and you have a very stable, viable solution.

Also you get 4 hour priority call outs on the SAN (so if the hard drive goes they have 4 hours to fix it and they do stick to it) which I find absolutely brilliant.



M.
 
m4cc45 said:
I'd say it was easily comparable to any other SAN out there. Chuck some fibre channel disks in there and you have a very stable, viable solution.

Also you get 4 hour priority call outs on the SAN (so if the hard drive goes they have 4 hours to fix it and they do stick to it) which I find absolutely brilliant.



M.
Click to expand...

I don't doubt support is very good, all our server kit is HP, but for advanced features the EVA comes nowhere near to EMC or Netapp (not that I'd expect it to, it's a whole load cheaper). We have all three deployed and we'd use the HPs more if they did the job as well as the others because they are so much cheaper...
 
What extra features are you after? I mean you can have replication between SAN's, iSCSI, etc. but they are available for a cost rather than stock.



M.
 
m4cc45 said:
What extra features are you after? I mean you can have replication between SAN's, iSCSI, etc. but they are available for a cost rather than stock.

M.
Click to expand...

Well when it comes down to it a lot of vendors offer those features, just some are much better than others. The EVA has a lot of faults, there are many different arrays i'd take over the EVA.

The whole Virtualisation thing is not as good as people think. Many admins have come from a background of control and having this taken away by an array that wants to decide where to put things is uncomfortable. When you have lots on a box this is not good, especially without any form of dynamic tuning.

Other arrays allow the admin to dictate where to put LUNs, on which exact disks. With the EVA my performance is affected by other apps using the same disk group - people need to give the applications predicable performance levels and this isn't possible in this situation. Only way to guarantee performance levels is to put it all on its own disks group - this is expensive, a lot of people are quite happy to share as long as they can pick, choose, limit, and have control if something does need moving.

Then there is the fact the EVA only supports restriping, not concatination. Painful for applications that are worldwide 24hour, someone in a timezone is going to get crud performance for a few hours.

What else? oh yeah - No method for destaging cache to disk in the event of a long power outage - after the new york blackouts a lot of people lost data. Is it too hard to destage to disk?

The need for a server for management, snaps, clones.

The fact that proactive hot sparing and per-disk-group hot spares each require 2x the largest disk size reserved in each group. This leads to half the damn array being full of spares, especially where people have made lots of groups to keep application (Exchange, SQL, etc) separate.

Smallest disk group is 8 disks. So if I wanted just a small amount dedicated of space for my exchange logs (like 4 disks) i'd have to buy 8 minimum. It's hard enough explaining to procurement why I need to keep exchange on dedicated spindles let alone buy 8 spindles just for for 100GB of logs! The alternative is to let the EVA do whatever and put it with other data but this means poor performance and is against MS best practice.

Same thing about FATA drives. If I just wanted a pair of large, cheap, FATA drives for some development scratch area in i'd have to but 8 because it's against HP best practices to mix the size and speed of drives in a group.

Well, that's some of what's wrong with the EVA. Other arrays don't suffer from any of this really.

Speaking about the EMC Clariion it has Quality of service tools to allow me to prioritise LUN so if I did have to restore a 1TB lun from somewhere I can make sure this wont affect my important luns, I can move Luns around inside the box (for example I can move a LUN from my expensive Fibre drives onto my cheap SATA drives), I don't need a server, LDAP integration, I can have a FC/iSCSI combo box (no HP equivalent without protocol coverters $$$$) and finally the equivalent EMC boxes has it beat on the hardware spec (memory, CPU, etc).
 
Sums up the problem with EVA nicely really, there are others, no integration with vmware in the same way EMC and netapp do (yet anyway, HP will have to come up with it in the end I feel) is a big one for me.

but...EMC is in another price bracket, I like the EVA still for when I need loads of cheap storage, we use it for our backup platforms, it's a great way to provide 25TB of data at a time at reasonable cost. I don't need shapshots and the basic replication is fine for my needs. I don't need iSCSI or lots of seperate disk groups for different apps, I need 48FATA drives presented as a LUN. For that, not much (not much with a recent reputation and reliable support) comes close to the price point.
 
You must log in or register to reply here.
Back
Top Bottom