RAID Controllers - Stupidly Expensive for what they are
-
@scottalanmiller said:
@Dashrender said:
not sure why anyone would want a Windows one.
SMB3, for example? DFS? ReFS? Management as part of the existing Windows infrastructure.
If you wonder why anyone would want Windows here, do you feel that Windows doesn't make sense ever or just for file servers?
Just for remote storage - Is SMB3 really getting any traction?
-
@Dashrender said:
Just for remote storage - Is SMB3 really getting any traction?
It's bit in the HyperV space for backing HyperV. It's got some benefits for normal users. I think that basically everyone on Windows file servers is on it now (that are on current Windows Servers.)
-
Search is horrible here, perhaps tagging will help with sorting things.
-
@Dashrender said:
Search is horrible here, perhaps tagging will help with sorting things.
Yes, you are very bad about tagging your posts. I have to edit every one to add tags after you post!
-
@scottalanmiller said:
@Dashrender said:
Search is horrible here, perhaps tagging will help with sorting things.
Yes, you are very bad about tagging your posts. I have to edit every one to add tags after you post!
Am I alone in this?
-
@Dashrender said:
@scottalanmiller said:
@Dashrender said:
Search is horrible here, perhaps tagging will help with sorting things.
Yes, you are very bad about tagging your posts. I have to edit every one to add tags after you post!
Am I alone in this?
Mostly.
-
-
Edit: Let's try this again...
I set up a Windows Server 2012 File Server Role as a Failover Cluster (it uses SMB 3 for transparen't failover). The transparent failover worked beautifully. We had a server keel over and release the magic smok and our end users barely noticed the blip as everything failed over to the other server in less than 5 seconds.One other benefit of a modern (Windows) file server is that Deduplication is relatively easy to set up, and on that same File Server, we were getting ~30% deduplication)
NB: This was the original post... skip it if you want.
I set up a Windows Server 2012 File Share Failover Cluster that was backed by a SAN (hush, @scottalanmiller ) and got to experience the SMB3.0 transparent failover a few times...when the file server randomly rebooted a few times. It...was... beautiful (as well as the Deduplication! ~30% savings on a 2TB). -
Uh - maybe I'm confused - I didn't think SMB was a SAN protocol?
-
@Dashrender You missed the part where it was served up from a Windows 2012 File Server... This is what happens when I try to talk about too many technologies in once paragraph, lol... I'll go edit that post for clarity.
-
@dafyre said:
I set up a Windows Server 2012 File Share Failover Cluster that was backed by a SAN (hush, @scottalanmiller ) and got to experience the SMB3.0 transparent failover a few times...when the file server randomly rebooted a few times. It...was... beautiful (as well as the Deduplication! ~30% savings on a 2TB).
What does backed by a SAN mean here?
So you have a Windows Server 2012 box that is sharing via SMB3.0 storage that is on a SAN - OK... the SMB3.0 is providing file services to end users?
-
@Dashrender said:
What does backed by a SAN mean here?
SAN backing is a standard term for what is sitting "behind" the storage that you see. The backing is lower in the stack, heads are higher in the stack. So this would be a one or more Windows file server head with SAN backing.
-
@Dashrender said:
So you have a Windows Server 2012 box that is sharing via SMB3.0 storage that is on a SAN - OK... the SMB3.0 is providing file services to end users?
Correct
-
@Dashrender said:
So you have a Windows Server 2012 box that is sharing via SMB3.0 storage that is on a SAN - OK... the SMB3.0 is providing file services to end users?
Yeah... Sorry for the incoherent babble I wrote. That hurt my head when I went back and re-read it... Post fixed, lol.
But yeah, that's what we run our home-folder redirection on. When a server blips or just flat out dies, the share moves to another member of the cluster and the users are none-the-wiser. 8-)... It has saved our bacon a couple of times.
-
The biggest question would be.... why are servers blipping?
-
@scottalanmiller said:
The biggest question would be.... why are servers blipping?
Dying Power Supply... Dying RAM.... Dying CPU... server dying of old age (these servers were ~8 years old-or better). It finally blipped it's last bleep a few months ago, lol.
-
@dafyre said:
@Dashrender said:
So you have a Windows Server 2012 box that is sharing via SMB3.0 storage that is on a SAN - OK... the SMB3.0 is providing file services to end users?
Yeah... Sorry for the incoherent babble I wrote. That hurt my head when I went back and re-read it... Post fixed, lol.
But yeah, that's what we run our home-folder redirection on. When a server blips or just flat out dies, the share moves to another member of the cluster and the users are none-the-wiser. 8-)... It has saved our bacon a couple of times.
Wouldn't DFS do this as well? did SMB 3.0 solve a problem that DFS did not? - I'm asking in earnest.
-
@Dashrender Technically, this is a problem that DFS should have helped us with... In practice, we had a lot of issues actually getting the files to be synced. Was probably the SysAdmin's fault in that case, though, lol. (Points fingers at me).
I'm not actually sure how DFS responds when a server completely dies... We didn't stay with it long enough.
-
@dafyre said:
Dying Power Supply... Dying RAM.... Dying CPU... server dying of old age (these servers were ~8 years old-or better). It finally blipped it's last bleep a few months ago, lol.
Wouldn't replacing the server with something stable instead of putting in a nice SAN and multiple dying servers have fixed that for cheaper?
-
@Dashrender said:
Wouldn't DFS do this as well? did SMB 3.0 solve a problem that DFS did not? - I'm asking in earnest.
Yes and no. DFS is meant to sort of do that but does it in a very different way and often does not work. DFS is a bit flaky and certainly not designed to be an HA solution.
SMB3 is much more enterprise for high reliability.
