Hi All;
I tried to search the archives on this but the service is "temporarily
unavailable"...
We lost a FCA in our EMC Symmetrix on Saturday and now (after rebooting all the
nodes) nothing can connect to the TSM server. This is the error I'm seeing in
the dsmerror.log:
01/06/03 07:56:47 TcpOpen: TCP/IP error connecting to server.
01/06/03 07:56:47 sessOpen: Failure in communications open call. rc: -50
01/06/03 07:56:47 ANS1017E Session rejected: TCP/IP connection failure
NO IPA's were changed, and I'm not sure what exactly to check on the TSM server
- though a quick check at "q status" looked normal. Any ideas?
Thank you;
Theresa Sarver
>>> jpendergast AT WATSONPHARM DOT COM 12/19/02 08:58AM >>>
Questions for you to ponder:
Are you restoring a TSM server, and then your Application server, or are your
restoring them both to the same box.
If item 1 is "one box", Do you already have the TSM server software sitting
dormant on the Application server.
Are you using DRM?
Have you already attempted a dry run on your in-house systems?
Should you add a recent backupset to your list of "Things to Take"?
Your list looks fairly complete. When we ran our DR test at Sungard, I had
forgotten the Atape drivers. If you install the Atape drivers on your server
before your mksysb, they will already be on the server.
We ran two different style tests on two different servers:
First one was a backupset restore. (Offers a quick and accurate restore, but
does not work well if a TDP is involved.)
1) Boot box with our sysback tape, and restore rootvg from sysback.
(Your routine will include the mksysb tapes and the savevgs)
2) Reboot and use the TSM client to restore the backupset using the overwrite
"ifnewer" option.
3) Reboot and test the application and we are done.
Second method, restore TSM server and application to the same box:
1) Same as other routine, boot box and load rootvg - remember to reboot
2) Create VG, LV, and FS as needed to load TSM database and logs
3) Load the DRM planfile from the floppy and expand it.
4) Adjust TSM configuration files as necessary for the change in environment (ie
tape)
5) Execute DRM script "RECOVERY.SCRIPT.DISASTER.RECOVERY.MODE"
6) When the DRM scripts asks for the dsm console to be started:
a) double check the dsm.opt and dsm.sys files on the client (ie change
server name)
b) start the TSM server if not already started
c) remove the drives/libraries
d) add the corrected library/drive
e) start the dsm console (the console allows you to see tape requests on
the server)
f) press return on the first session to complete the DRM script
7) restore the OS and application using the overwrite "ifnewer" option - a
reboot after might be good for safety
8) restore the application data using the TDP software
Theresa Sarver <tsarver.IFMC AT SDPS DOT ORG> on 12/18/2002 12:34:42 PM
Please respond to "ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>
To: ADSM-L AT VM.MARIST DOT EDU
cc: (bcc: Joseph Pendergast/Corona/Watson)
Subject: Disaster Recovery Test
Hi all;
I'm getting ready to preform a Disaster Recovery test (with Sungard) and I was
wondering if anyone would be interested in reviewing my steps and letting me
know of anything I may be missing. - Thanks in advance!
THINGS TO TAKE:
mksysb (and savevgs)
AIX 4.3.3 cdroms
Atape drivers
DLT tape drivers
Application Software (Stars/Sql_backtrack/OBSI)
Database software (Sybase)
Tivoli Software
All license Information
Steps to Restore:
1) Restore mksysb
2) Install ATAPE/DLT drivers
3) Restore savevgs
4) Restore TSM database
5) Restore "Export Node" data
6) Restore OS data (not sure if this step is necessary?)
7) Install Application/Database Software (not sure if this step is necessary?)
8) Restore Sybase [raw] data
9) Check network connectivity
Also, our environment has 2 internal DLT tape drives, I will only have 1
external DLT drive during the test - will Tivoli have a problem [restoring] with
this?
Any advice or comments would be much appreciated - thanks again!
Have a happy holidays;
Theresa
|
