HELP! File Server 2008 R2 hard drive problems

[mcoyne]

Hi everyone,

 

As for what the original problem was, still none the wiser. I'm going to guess at a corrupt VMDK file. Anyhow - one week later and we have a new whizzy duper file server in production, so as far as the end users are concerned, problem solved and now the pressure is off.

 

NOT A FUN WEEK!!!

 

Thank you everybody for your assistance and I definitely have that Friday feeling...

 

Cheers,

 

Mike

[zag]

Yep you did the right thing. I always try to design systems that have a quick failover. I personally run 2 identical file servers all the time. The "spare" just sits there in case the primary one goes down.

 

File servers are too important to centralize all your storage on a single piece of hardware such as a SAN in my opinion. Much better to distribute over local storage from experience.

[mcoyne]

Hi Zag,

 

Thank you for your thoughts and I'm glad you think so! Yes, the SAS box was put in by Capita under the Building School's for the Future project. I think they did a pretty good job on the whole but they were trying to achieve too much with not enough money in my opinion, so there were a few shortcuts.

 

I like your idea about the spare file server! I've never used it, but could you keep the second one current by use of DFS replication? It's something I've been meaning to dabble with at some point!

[zag]

Hi Zag,

 

Thank you for your thoughts and I'm glad you think so! Yes, the SAS box was put in by Capita under the Building School's for the Future project. I think they did a pretty good job on the whole but they were trying to achieve too much with not enough money in my opinion, so there were a few shortcuts.

 

I like your idea about the spare file server! I've never used it, but could you keep the second one current by use of DFS replication? It's something I've been meaning to dabble with at some point!

 

Yep replication would give you a nice option, but brings its own complications and issues sometimes. Imagine a ransom-ware attack scenario and you see how instant replication can fall down.

 

Personally I try not to over engineer these things. I just run a scheduled xcopy /mirror each half term.

 

We of course have other backups, but this gives the peace of mind that if the file server ever just dies then I could switch over in a few minutes without any stress at all.

[mcoyne]

All very very good points Zag - I like your style! Yes, I think sometimes we can over engineer just because we can, but often simpler is more effective. Very good food for thought :-)