+100 pts for appropriate symbol use. -- ME2
On Fri, Sep 24, 2010 at 8:42 AM, Jonathan Link <[email protected]>wrote: > +*∞* > > On Fri, Sep 24, 2010 at 11:38 AM, Ken Schaefer <[email protected]>wrote: > >> I think this thread (like all the previous ones) has gone on long enough. >> >> We are not asking you questions because we want answers. Frankly, I think >> most of the people here no longer care - you've used enough time as it is. >> >> *You* need to work out what your *requirements* are. Not what you're >> 'unhappy' about. Or what you 'think' you need. You need to find out what the >> business needs, in order of priority. >> >> For example you state that you need to be back up and running within 3 >> days. OK - a SAN is not going to help with that. Only a *recovery* system >> can help with that. That means some way of replacing your tape drive (if you >> are worried that you'll lose that), and a way of getting your tapes back, >> and a way of restoring. All within 3 days. That's called your RTO: Recovery >> Time Objective. >> >> The next thing to consider is your RPO - Recovery Point Objective. How >> much data can you afford to lose? One day? Two days? A week? Again *you* >> need to figure this out. And again, a SAN will not help you with that. >> >> The only thing a SAN is going to do is help you avoid a recovery scenario. >> But you haven't stated *any* requirement whatsoever about this. Is the >> business happy to pay $30k to ensure that they only have a disaster once >> every 10 years? Or would they prefer to suffer a disaster once every 5 >> years, but by spending $30k on a tape library, they can be up and running >> again in 3 days? This is what *you* need to find out. Then you can work out >> what you need to buy. >> >> It doesn't matter how big or small your environment is you need >> requirements. My environment is going to be ~4000 Wintel servers in >> Production alone, I suspect yours will be smaller unless the carpet business >> significantly picks up. Our requirements from the customer and internally >> run to many hundreds of pages - probably over a thousand pages now. Even my >> home network (where I have about 10 VM servers) has requirements. Otherwise, >> you are just going to be either (a) bothering people with questions forever >> or (b) p*ssing money up a wall on stuff you don't need. >> >> If you want help documenting what you need, then please ask for help on >> that. Please stop asking for advice on SAN vendors until you've worked out >> what your requirements are, and you think you've found a good fit and what >> other people's experience with that particular piece of kit. >> >> Cheers >> Ken >> >> -----Original Message----- >> From: John Aldrich [mailto:[email protected]] >> Sent: Friday, 24 September 2010 11:26 PM >> To: NT System Admin Issues >> Subject: RE: SAN question >> >> We are running DFSR, but only for redundancy. All clients map to a >> physical machine and drive, as we had some issues with DFSR not staying >> synchronous, even over a GigE connection. This was mainly due to running out >> of room on the disk for replication (due to multiple copies of large files >> being stored >> --- since corrected.) >> >> >> >> >> >> -----Original Message----- >> From: [email protected] [mailto: >> [email protected]] >> >> Sent: Friday, September 24, 2010 11:06 AM >> To: NT System Admin Issues >> Subject: RE: SAN question >> >> Also do you use DFS? If you do, NAS units don't work. The volumes must >> be mapped to Windows servers as local drives (meaning ISCSI or DAS) >> >> HECK, running two servers with appropriate DAS running DFS/Replication >> would give you redundancy.. There are tons of ways to slice this without >> going to a SAN and spending that money unless your REQUIREMENTS dictate >> specific features that only SANS require. >> >> You can get two cheap Drobo or Synology boxes that support AD, SMB, CIFS, >> ISCSI (mini sans basically) 3 to 5 TB depending on raid and size of drives >> for 1/3rd the cost of a SAN. Synology and Drobo do replication between each >> other, you could use ISCSI and do DFS replication one to each server for >> redundancy, or have one online and replicate to the other for backups. >> >> >> -----Original Message----- >> From: John Aldrich [mailto:[email protected]] >> Sent: Friday, September 24, 2010 10:22 AM >> To: NT System Admin Issues >> Subject: RE: SAN question >> >> Well, my (admittedly limited) understanding is that at the low-end SANs >> have a lot of overlap with NAS and that they are almost interchangeable. I >> want some sort of separate machine to get the "file server" role off the >> DCs. >> Maybe that means a NAS, maybe it means a SAN, maybe it means a server >> with DAS running Windows Storage Server. At this point, I'm not really sure >> what the best money would be. Whatever we get, I want it to be expandable so >> that as we (hopefully) grow, we can add more storage as needed. >> >> I do like the idea of having tape to back up whatever we have. If we're >> going to have email in-house, we're likely to end up with at least a couple >> terabytes of data in the long run, so whatever archival backup we end up >> with is likely to need to be a library, instead of just an on-board tape >> drive. >> >> >> >> From: Kevin Lundy [mailto:[email protected]] >> Sent: Friday, September 24, 2010 9:12 AM >> To: NT System Admin Issues >> Subject: Re: SAN question >> >> And absolutely none of that requires a SAN. Especially for your data set >> size. >> >> Why do you think you need a SAN? versus NAS? versus well architechted >> DAS with decent tape? >> On Fri, Sep 24, 2010 at 8:37 AM, John Aldrich < >> [email protected]> >> wrote: >> I want to ensure that the data integrity remains intact, even if it takes >> a couple days to recover. This is business-critical data, although we could >> live without it for a couple or three days, it would be very difficult and >> time consuming to recreate much of the data on the servers. For this reason, >> I want redundant disks, network, controllers, etc. >> I believe I previously mentioned that my CEO told me we could live with >> taking up to 3 or 4 days to recover the data, but after that, it would be >> problematic. Personally, I'd like to get it down to under 48 hours to >> recover (not 4 business days, 48 actual hours.) That's why I want redundant >> controllers or if I can't get redundant controllers on the storage appliance >> itself, I want redundant storage appliances, such that the data itself is >> redundant. >> I would not like to have to go to the CEO and tell him "sorry, we lost the >> data because the system crashed and we had no backups." Theoretically, I >> could have one "appliance" and a tape library and be good, but I'd prefer to >> have it a *little* more robust than that. >> >> >> >> ~ Finally, powerful endpoint security that ISN'T a resource hog! ~ >> ~ <http://www.sunbeltsoftware.com/Business/VIPRE-Enterprise/> ~ >> >> --- >> To manage subscriptions click here: >> http://lyris.sunbelt-software.com/read/my_forums/ >> or send an email to [email protected] >> with the body: unsubscribe ntsysadmin >> >> > ~ Finally, powerful endpoint security that ISN'T a resource hog! ~ > ~ <http://www.sunbeltsoftware.com/Business/VIPRE-Enterprise/> ~ > > --- > To manage subscriptions click here: > http://lyris.sunbelt-software.com/read/my_forums/ > or send an email to [email protected] > with the body: unsubscribe ntsysadmin > ~ Finally, powerful endpoint security that ISN'T a resource hog! ~ ~ <http://www.sunbeltsoftware.com/Business/VIPRE-Enterprise/> ~ --- To manage subscriptions click here: http://lyris.sunbelt-software.com/read/my_forums/ or send an email to [email protected] with the body: unsubscribe ntsysadmin
