Hey, we're looking at maybe migrating our infrastructure over to a set of blades (10 bay enclosure, 8 servers online with two spares at the ready), and I'd like to talk with someone who's worked with them in a production environment.
Some questions I have are:
1. Dell mentions that you can "Provision" a blade slot for a particular OS configuration for easy restore in the event of failure. How does that work? Do you keep a preconfigured 2003 server image on hand? Does it take "snapshots" of that slot's hardware at predefined intervals?
2. How does the OS see the SAN? Is it seen as a network drive or does it appear more inline with how your box would see a SCSI controller and drive?
3. How do you manage filesystems on the SAN? Does it support multiple partitions? NTFS? FAT? Proprietary? Managed through special proprietary software or through Windows' Disk Management Console?
4. MS Exchange server. I assume that you'd store the database on the SAN rather than on the blade itself, but what about recovery in the event that blade fails? This ties back to question #1. If we're dealing with snapshots, that's cool unless they're over 30 days old, in which case you could have some Kerberos problems with that machine on the domain. Not hard to fix, still, but a pain in the ass any way you look at it.
5. What about automatic failover? Can you provision a spare box to keep itself synched, configuration-wise (not file system-wise, because if a virus is what brings your server down you sure as hell don't want to have a backup of that virus
that will automatically take over in the event the first box dies or just goes crazy? If it does, will it shut down that other box in order to prevent network problems?
If anyone's fairly nearby (southern california, I'm in Rancho Cucamonga) and has such an environment set up, I'd love to check out your setup and talk to you first hand about the pros and cons of using Blade servers over standalones.
Thanks!
Jason
Some questions I have are:
1. Dell mentions that you can "Provision" a blade slot for a particular OS configuration for easy restore in the event of failure. How does that work? Do you keep a preconfigured 2003 server image on hand? Does it take "snapshots" of that slot's hardware at predefined intervals?
2. How does the OS see the SAN? Is it seen as a network drive or does it appear more inline with how your box would see a SCSI controller and drive?
3. How do you manage filesystems on the SAN? Does it support multiple partitions? NTFS? FAT? Proprietary? Managed through special proprietary software or through Windows' Disk Management Console?
4. MS Exchange server. I assume that you'd store the database on the SAN rather than on the blade itself, but what about recovery in the event that blade fails? This ties back to question #1. If we're dealing with snapshots, that's cool unless they're over 30 days old, in which case you could have some Kerberos problems with that machine on the domain. Not hard to fix, still, but a pain in the ass any way you look at it.
5. What about automatic failover? Can you provision a spare box to keep itself synched, configuration-wise (not file system-wise, because if a virus is what brings your server down you sure as hell don't want to have a backup of that virus
that will automatically take over in the event the first box dies or just goes crazy? If it does, will it shut down that other box in order to prevent network problems?If anyone's fairly nearby (southern california, I'm in Rancho Cucamonga) and has such an environment set up, I'd love to check out your setup and talk to you first hand about the pros and cons of using Blade servers over standalones.
Thanks!
Jason
Facebook
Twitter