Storage and backup solution (iSCSI SAN) - reassign initiator

We are planning a new storage and backup solution. This is our proposed layout. All servers running Windows Server 2003.

1 iSCSI SAN (Promise VTrak M300i (SATA))
1 Gb switch (3Com 3824)

HP Ultrium 448 200/400 GB (for off-site backup)
Veritas Backup Exec 10d + Advanced Disk Backup Option

SAN is intended to be used by two of our servers, so the storage will be split up into three volumes. One for the file server, one for the email server, and the third for a disk backup solution.


If one server goes down, how easy is it to transfer the corresponding volume to another server? (Reassign the initiator?)

Is layer 3 switching recommended or is layer 2 sufficient?

What types of SATA HDDs is recommended?

Anyone got any experiences with Promise VTrak?

Any other comments/suggestions?
Who is Participating?
IPKON_NetworksConnect With a Mentor Commented:
Generally speaking, it is a very quick task to reassign a volume to another host. Point and click and posisbly a reboot (if using basic disks in Windows) and the volumes will be present. If you are using a flash copy type backup to the spare volumes for fast recovery then you will need to extra areas of SAN.

1. Live file server
2. Live Mail server
3. backup snap/flash copy area for file server
4. backup snap/flash copy area for mail server

You would then perform the snap copies of the two live volumes to the empty 'volumes' which would provide you with a instantly available recovery process. Basically, you do the following (assuming the volumes are just DATA and not OS)

Hardware/server failure:
1. Rebuild a new server (either new hardware and OS install, or re-install onto existing hardware)
2. On the SAN management console, pick up the HBA's for your new server (if they have changed)
3. Point the new server to the existing data volume and reboot.

Data volume failure:
1. In SAN management console, remove the volume from the host
2. allocate backup volume to host
3. reboot server to make active.

Windows stored ALL share information on the C:\ within the registry. You MUST restore this to any new OS installation if you want to keep your file shares in the 1st recovery process.

Any binaries that are stored on the C:\Program Files folder will be lost in the first instance as well. These would need to be reinstalled
If you store the binaries on the SAN then you need to capture a System State onto your Data Volume before doing the snap volume copy to ensure you can recover the registry etc to its prior state.

MS Exchange (if using this) is a little picky about backup/recovery.... I'm being nice btw! You should always try and do an offline backup in my experience. This removes a lot of issues when doing a recovery. This is especially true for full server recoveries. This way you won't need to try and bring your IS files back into sync after you recover using either method above.

Hope this helps
pgm554Connect With a Mentor Commented:
I would go with clustering,spend the extra money and then you won't have to deal with worry of being down for extended periods.

As for disks WD Raptors,they are as expensive as SCSI,but the HDA's are built for longevity.
The failure rate for regular drives is quite high(Maxtor being quite bad right now).
Seagate bought them ,so let's see if the QA improves.
risoyAuthor Commented:
What do you mean regarding storing "share information on the C:\ within registry"? Is D:\ "safe"?

Backup of the operating system and program files are not neccesary in my opinion. If the server crashes we will replace it with another one. The affected services can easily and quickly be transferred to another server while we replace the original.

Have a look at this link for a proposed structure:
tellawiConnect With a Mentor Commented:

I guess what it means by "share information on the C:\ within registry" is windows systems data. As you know, in iscsi, you must have a local installation of windows, you cannot install windows in a external storage using iscsi like normal SAN solution.

One thing to add is in iSCSI, you cannot map the same iSCSI drive (or LUN) from two servers at the same time,  you will risk data integrity. Also put in mind that when you write data into iscsi drive, data will not be written to the disk immediately.
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

All Courses

From novice to tech pro — start learning today.