Meir,
I think you are referring to HP SFS which is a product based on
Lustre.
I should point out it is possible to create filesystems of equivalent
sizes as the Lustre
codebase is common.
VMS logical name tables were an artefact of the operating system, akin
to environment
variables in UNIX/Linux. They had the capability of including search
lists such that an
access to an item by logical name would search for the item according to
the sequence
in the search list. Thus if an item were only to be found at the second
location in the
list, the OS could still find it.
Such a capability had less to do with the filesystem than with the
operating system,
so its not completely a question of whether any lustre filesystem would
handle it.
Lustre (and consequently SFS) is a parallel filesystem - parts of the
data are stored in
stripes on OSTs in parallel. Acces to files uses the normal POSIX apis,
so no code changes.
Lustre configurations can be architected to have redundant paths to
storage so that client
nodes can access data from alternate servers in the case of server
failure. You may have
seen other notes on this topic in lustre-discuss. All HP SFS
configurations have redundant
alternate servers. Redundancy in the backing storage can help cater for
disk failures,
and Lustre can avail of this just as any filesystem could.
- Tom
---
Tomas Hancock, Hewlett Packard, Galway. Ireland +353-91-754765
The contents of this message and any attachments to it are confidential
and may be legally privileged. If you have received this message in
error you should delete it from your system immediately and advise the
sender. To any recipient of this message within HP, unless otherwise
stated, you should consider this message and attachments as "HP
CONFIDENTIAL".
________________________________
From: lustre-discuss-admin@lists.clusterfs.com
[mailto:lustre-discuss-admin@lists.clusterfs.com] On Behalf Of M
Sent: 23 November 2005 13:19
To: lustre-discuss@lists.clusterfs.com
Subject: [Lustre-discuss] a question about logical names
i ask this question because i read that "hp" has created a grid file
system ("fsf") that is based on "lustre", but has twice
the capacity of lustre from 250tb to 500tb.
in the past digital/hp had a file system called: "vms" , and
"vms" had
logical names and logical name tables.
the idea was that if a cluster of disks does not work, the program that
"works" with logical name like: "my_disks1"
is going to search the file in another cluster of disks, and in this way
allows hot backup,
you did not need to enter the source code and change the place of local
file inside - just at the logical name.
and even at definition time you could create a few names for the same
logical name, so in case of crash - the
program could find the file in a different cluster of disks.
i need to know if "lustre" / "fsf" have this ability?
thank you
meir abramson
meirab@bigfoot.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://mail.clusterfs.com/pipermail/lustre-discuss/attachments/20051123/7a1e682b/attachment.html