|
Re: [Networker] True Disaster Recovery
2006-10-03 16:28:56
On Oct 3, 2006, at 12:57 PM — 10/3/06, Albert Eddie Contractor AFRPA
CIO/IT wrote:
I read (with interest) the thread about disaster recovery. (Recovering
tapes on Server B if Server A dies and you can not change Server B
name
because it is in use as a full time server)...
Our disaster plan has to be a little more "NEW ORLEANS" (what if the
site is completely wiped out!?)
I love reading other people's disaster plans even while I am refine my
own. Do any of you have disaster plans you could share with me?
Of particular interest is complete SERVER ROOM disaster (aka no other
server in room can be used)....
Some of the best plans (on a budget) that I have read have included
LEASING a server that meets the H/W & S/W requirements on a month by
month basis while you rebuild your server room with PURCHASED H/W.
Would love to hear your thoughts.
Our DR plan here at Temple U. is to classify each of our servers into
different levels such as 0, 1, 2. Level zero servers get top priority
for recovery in the event of a major disaster, al la Katrina. The
higher the level number, the lower the priority. Level zero servers
must have a duplicate server at one of our two DR sites. One is owned
by Temple and is on Temple's main campus. The other site is rented
from Sunguard, but it is still here in Philadelphia. The Sunguard
site is where our old mainframe iron's DR stuff is located. We also
have a small LDAP farm, DR web server, and email server (for
executive board members) at Sunguard.
Our on campus DR site is about 1/4 mile from our main data center.
Our only NetWorker server is located in the on-campus DR site. We
also keep some small administrative servers there that are DR level 0
and level 1 servers in our data center.
We have quarterly rehearsals that involve random server audits to
ensure that fail over to our DR stuff works. As far as the DR stuff
that's backed up by the NetWorker server that I manage (which is the
majority of our data), we have a 100% success rate in meeting DR
requirements. Our mainframe DR is handled outside of NetWorker and is
not within my purview ... yet.
We ship mainframe data off-site six days a week. All data that's
backed up via NetWorker sits on tapes in our Sony PetaSite and those
tapes are never removed from the tape library, nor do we copy them.
This is because the library is already outside the building where the
primary copies of the data are maintained. Plans are also in progress
to do a major revamp of how our mainframe data (e.g., student grades,
payroll, etc.) are backed up, so our mainframe DR strategy is likely
to change.
Our NetWorker server consists of a storage node and a main server,
both running NSR 7.2.1. The server has a huge Sony PetaSite connected
to it, which is on its on fibre SAN. Due to the high cost of
purchasing our Sony tape library, we have not taken steps to
replicate it anywhere at all. If our NetWorker server goes up in
smoke though, we lose no production data because 100% of the
production data is in a different building. Our data center and on-
campus DR center are well protected structurally from environmental
disasters and Temple can provide its own power in an emergency. The
DR site also is guarded by a bunch of automated protections from fire
and unauthorized intrusion, as is our primary data center. As a
result, the level of exposure to business discontinuity if our
NetWorker server died is very remote.
We are shopping for a new Chief Information Security Officer. Once a
new CISO is hired, it would not surprise me at all if having DR plans
for our NetWorker server enters the picture, but I think management
will be shellshocked at the cost to protect our NetWorker server and
our main tape library. In terms of hardware cost, our NetWorker
server and its tape library are the most expensive server Temple
owns, by far and I would say that licensing it with EMC probably
ranks second or third, so I would be surprised if we actually take
any real steps to protect our backups from a Katrina style disaster
due to the high cost, but I might be wrong and I do at least expect
this question to get on the radar screen at some point this coming year.
We also have one storage node and and old tape library still online,
but out of production. We are seriously considering using that
hardware for our mainframe backups outside of NetWorker (lack of IBM
mainframe support), but that issue is still undecided.
To sign off this list, send email to listserv AT listserv.temple DOT edu and type
"signoff networker" in the
body of the email. Please write to networker-request AT listserv.temple DOT edu
if you have any problems
wit this list. You can access the archives at
http://listserv.temple.edu/archives/networker.html or
via RSS at http://listserv.temple.edu/cgi-bin/wa?RSS&L=NETWORKER
|
|
|
