Network Applicance V's EMC

Does anyone have experience with both Netapp SANs and EMC SANs,
supporting MS SQL 2000. If so, what's your opinion of one versus the
other?
We're currently using a netapp f825c and are wondering how EMC
compares.
Tom Chute
do you mean network appliance "NAS" ?
Greg Jackson
PDX, Oregon
|||FYI: Yes, I do mean SAN. We're using a fiber channel to connect an
825c to an MS SQL Server. We have a single database split across two
filer heads.
|||No, I mean SAN. We're using a fiber connection to connect a SQL Server
to an 825 cluster. We have a single database split across two filer
heads.
The problem is that we've outgrown the 825c and need to upgrade. I like
Netapp, but my management wants me to look at EMC. I have no experience
w/ EMC, and was wondering if anyone out there has. If they have, what
was their opinion of dealing w/ Netapp and using a Netapp SAN V's EMC.
Tom C.
*** Sent via Developersdex http://www.codecomments.com ***
Don't just participate in USENET...get rewarded for it!
|||tom_chute01@.yahoo.com (Tom Chute) wrote in message news:<b5d5c748.0410081554.41c2fb84@.posting.google. com>...
> FYI: Yes, I do mean SAN. We're using a fiber channel to connect an
> 825c to an MS SQL Server. We have a single database split across two
> filer heads.
Tom,
I just finished a very lengthy trial of NetApp FAS940's and EMC
CX700's. From a LUN performance perspective, EMC just edged out
NetApp mainly due to EMC's 8 GB of cache. Administratively, I liked
EMC better as they have a built in administrative console (Java based)
to manage multiple EMC systems and all the their functions from one
console. As you probably know, NetApp requires a different
application for just about everything they do.
Where NetApp pulls ahead is in the application support and replication
arenas. EMC does a bad job of managing Exchange and SQL replication
and snapshots. One of our main purposes for going with a large SAN
environment was for disaster recovery. The big question we had for
EMC is how do they bring back changed SQL and Exchange data from a
disaster site once the disaster is over? Their first response was to
bring over the entire database (SQL or Exchange). Our Exchange system
is over 700 gigs and we have SQL servers well over 300 gigs. Even
with a DS3 we're looking at several hours to replicate (our
replication site is 1000 miles away in a branch office).
EMC then came back with an incremental restore option that was
basically a log file rollup. Essentially the same thing as a manual
Exchange or SQL database restore. EMC just didn't offer anything
above what manual steps are required to restore databases. This may
not be an issue if you're only doing local snapshots. Both EMC and
NetApp require about the same amount of overhead for a LUN snapshot.
However, NetApp has just announced improvements to their LUN snapshots
that require less than 100% LUN overhead.
With NetApp's SnapManager for Exchange and SQL, I can perform an
incremental restore from my disaster site directly to my production
SQL databases or Exchange datastores without stopping services,
rebooting servers, and without user interruption. We internally setup
a FAS system in both locations and performed live tests with test
data. I can literally move to my disaster site, make as many changes
as I want, replicate back only the changed blocks of data back to my
production system and be working again in only minutes above the time
it takes to replicate the data back. And since NetApp does everything
on a true changed block basis, it takes me only minutes to recover
changes rather than hours or days to bring over the entire database.
The trade-off? NetApp costs twice as much as EMC. NetApp may make you
install a few more apps and administrativly it may not be as
centralized but we choose NetApp simply because it will allow us to
work with our applications better than EMC.
EMC is the monster of SAN's and maybe a Symetrics with SRDF might have
worked better but that was getting out of our comfort level for
administration and I don't think they have any different support for
enterprise applications.
Hope this helps,
JT
|||Hi
We are running EMC Symmetrics with SRDF (going A-SRDF as soon as we can) to
a remote site over 1Gb fiber. Pushing 0.5Tb an hour, capable of 19Tb an
hour. 128GB on the Cache Controllers.
Big environment, but we just have to have 99.999 uptime.
Once you start to look at large SAN's, 'go big or go home'.
Photons, Clarions, DMX's and local storage are being phased out and we will
then only be running a dozen EMC's.
We are happy and won't consider anything else.
Cheers
--
Mike Epprecht, Microsoft SQL Server MVP
Zurich, Switzerland
IM: mike@.epprecht.net
MVP Program: http://www.microsoft.com/mvp
Blog: http://www.msmvps.com/epprecht/
"JT Johnson" <jtjohnson@.faegre.com> wrote in message
news:b09ab426.0410110924.27bd276c@.posting.google.c om...
> tom_chute01@.yahoo.com (Tom Chute) wrote in message
news:<b5d5c748.0410081554.41c2fb84@.posting.google. com>...
> Tom,
> I just finished a very lengthy trial of NetApp FAS940's and EMC
> CX700's. From a LUN performance perspective, EMC just edged out
> NetApp mainly due to EMC's 8 GB of cache. Administratively, I liked
> EMC better as they have a built in administrative console (Java based)
> to manage multiple EMC systems and all the their functions from one
> console. As you probably know, NetApp requires a different
> application for just about everything they do.
> Where NetApp pulls ahead is in the application support and replication
> arenas. EMC does a bad job of managing Exchange and SQL replication
> and snapshots. One of our main purposes for going with a large SAN
> environment was for disaster recovery. The big question we had for
> EMC is how do they bring back changed SQL and Exchange data from a
> disaster site once the disaster is over? Their first response was to
> bring over the entire database (SQL or Exchange). Our Exchange system
> is over 700 gigs and we have SQL servers well over 300 gigs. Even
> with a DS3 we're looking at several hours to replicate (our
> replication site is 1000 miles away in a branch office).
> EMC then came back with an incremental restore option that was
> basically a log file rollup. Essentially the same thing as a manual
> Exchange or SQL database restore. EMC just didn't offer anything
> above what manual steps are required to restore databases. This may
> not be an issue if you're only doing local snapshots. Both EMC and
> NetApp require about the same amount of overhead for a LUN snapshot.
> However, NetApp has just announced improvements to their LUN snapshots
> that require less than 100% LUN overhead.
> With NetApp's SnapManager for Exchange and SQL, I can perform an
> incremental restore from my disaster site directly to my production
> SQL databases or Exchange datastores without stopping services,
> rebooting servers, and without user interruption. We internally setup
> a FAS system in both locations and performed live tests with test
> data. I can literally move to my disaster site, make as many changes
> as I want, replicate back only the changed blocks of data back to my
> production system and be working again in only minutes above the time
> it takes to replicate the data back. And since NetApp does everything
> on a true changed block basis, it takes me only minutes to recover
> changes rather than hours or days to bring over the entire database.
> The trade-off? NetApp costs twice as much as EMC. NetApp may make you
> install a few more apps and administrativly it may not be as
> centralized but we choose NetApp simply because it will allow us to
> work with our applications better than EMC.
> EMC is the monster of SAN's and maybe a Symetrics with SRDF might have
> worked better but that was getting out of our comfort level for
> administration and I don't think they have any different support for
> enterprise applications.
> Hope this helps,
> JT