Some “server” needs to host the global lock manager, but in the abstract, it
doesn’t matter which server does it – e.g., the GLM does not need to be
co-located with the resource. The GLM can even be distributed, with
responsibility for a particular lock’s management migrating to the node with
the most frequent requests for the given lock. That server’s identity
becomes an attribute of the lock.
IIRC, the VMS locks included a field that could be used as a serial number,
which was incremented each time an exclusive mode was released. This was a
simple way for cache managers to know when their cache was stale: For each
Read access to cached data, acquire the lock shared; if the serial number
has changed since the data was cached, refresh.
Putting the GLM in the disk array is an interesting idea, since it would be
implemented in hardware and accessed at SAN bus speeds, rather than in a
software server accessed over the LAN. The Lock, SN Check and Disk Read
could be combined in a single call, implemented in the device, where the
data is not transferred if the lock serial number is current. A single
mechanism of this sort could support both file level and block level
locking.
-----Original Message-----
From: Jamey Kirby [mailto:xxxxx@storagecraft.com]
Sent: Friday, February 22, 2002 2:45 PM
To: File Systems Developers
Subject: [ntfsd] Re: How to share storage device on logical block acce ss
level?
I suspect it is not as much the cache manager, but how the current file
systems use the cache manager. To support shared disks, you will need a
whopping cache consistency and locking mechanism; similar to oplocks.
However, oplock are controlled on the server and in a SAN the device itself
is a server.
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Geoff Clow
Sent: Friday, February 22, 2002 2:27 PM
To: File Systems Developers
Subject: [ntfsd] Re: How to share storage device on logical block acce ss
level?
How does NT’s cache manager impede a shared device file system? The FS
controls whether a file is cached, and when that cache is flushed for cache
coherency. One problematic case is VMM mapped files, but disallowing shared
write access to mapped files seems a trivial restriction.
-----Original Message-----
From: Smith, Joel [mailto:xxxxx@ntpsoftware.com]
Sent: Friday, February 22, 2002 10:37 AM
To: File Systems Developers
Subject: [ntfsd] Re: How to share storage device on logical block acce ss
level?
Why, do you suppose, is there no commercial cluster file system for
Windows? It there something in the design of the OS that precludes the
development of a true cluster FS (perhaps this is intimated by your last
paragraph) or is there simply no commercial market for such a thing?
-Joel
-----Original Message-----
From: xxxxx@veritas.com [ mailto:xxxxx@veritas.com
mailto:xxxxx ]
Sent: Friday, February 22, 2002 8:17 AM
To: File Systems Developers
Subject: [ntfsd] Re: How to share storage device on logical block access
level?
On 02/21/02, ““Anton Kolomyeytsev” ” wrote:
> Hi,
>
> I have a storage device (hard disk) that is accessable from more then one
> machine in the same time (let’s assume it’s a SCSI hard disk on shared
> SCSI
> bus). While the device is only readen everything is OK, but how can I make
> this device writable by more then one client in the same time? How to keep
> the file caches consistant? Any ideas? Anybody works with clusters here?
> File system drive will not work I need to use existing one (maybe modified
> with filter driver but I’d prefer not to do so…)
>
> Regards,
> Anton
>
> P.S. For now I do store the written data in the different space so the
> hard
> disk is “partially shared”. Only the data that was never touched by
> writing.
>
> —
I worked on “Digital Clusters for Windows NT” several years ago, and we
determined that there was no simple answer. Thus we developed the
“sequential sharing” model for shared SCSI disk usage, which MSCS still
uses today. At any given time, only one node is allowed to access a shared
disk.
Many other people have replied, but only Andy Champ has given you a clue
about the answer. The answer is - you can’t do it without a true cluster
filesystem, running on all computers sharing the disk, and communicating
with each other. Other operating systems have achieved this in the past
(the classic is DEC’s VMS Clusters, released in 1984). But as of today, no
commercial cluster filesytem has been released for NT.
You could not get the right results with a simple filter driver, and might
even have troubles with a full filesytstem driver too, since the NT cache
manager is also involved in writing to disk.
Good luck!
Carl Appellof
VERITAS Software Global Corporation
—
You are currently subscribed to ntfsd as: xxxxx@ntpsoftware.com
To unsubscribe send a blank email to leave-ntfsd-$subst(‘Recip.MemberIDChar’)@lists.osr.com
—
You are currently subscribed to ntfsd as: xxxxx@stbernard.com
To unsubscribe send a blank email to leave-ntfsd-$subst(‘Recip.MemberIDChar’)@lists.osr.com —
You are currently subscribed to ntfsd as: xxxxx@storagecraft.com
To unsubscribe send a blank email to leave-ntfsd-$subst(‘Recip.MemberIDChar’)@lists.osr.com
—
You are currently subscribed to ntfsd as: xxxxx@stbernard.com
To unsubscribe send a blank email to leave-ntfsd-$subst(‘Recip.MemberIDChar’)@lists.osr.com
—
You are currently subscribed to ntfsd as: $subst(‘Recip.EmailAddr’)
To unsubscribe send a blank email to leave-ntfsd-$subst(‘Recip.MemberIDChar’)@lists.osr.com</mailto:xxxxx>