on 06-09-2010 4:10 PM
We have installed MSCS. THe way it is installed is we have Four gorups created.We have Enqueue replication server created. We are based on NW 04S architecture
1) CLuster Group
Resources on this Group are :-
IP address
Network Name
Quorum Disk
2) MSDTC Group
Resources on this Group are :-
MSDTC IP
MSDTC Network Name
MSDTC Physical Disk
MSDTC Transaction CO-rdinator
3) SAP Group
Resources on this Group are :-
Physical disk for \Use\sap\<SID>
SAP <SID> Instance 00 Resource (SAP Resource)
SAP <SID> Service 00 Resource (Generic Service)
SAP <SID> Instance 01 Resource (SAP Resource)
SAP <SID> Service 01 Resource (Generic Service)
Virtaul Network Name
4) SQL Data ( SQL DB) Group
Resources on this Group are :-
Physical disk for DB datafile
Physical disk for Log files
Physical disk for TempDB
SAP gateway
Virtual Network IP
SQL Server
SQL server Agent
My question,
1) Suppose , Both my SAP and SQL resources are running on Node 1. when I move SAP GROUP from Node 1 to Node 2, when it happens all end users get disconnected. Does it supposed to disconnect all end users when I move SAP GROUP from one node to other Node 2 ?
I thought failover mechanism should be transperent to users i.e when failover happens following things should happen
1)End Users should get "hour glass" .
2)All lock are preserved (enqueue)
3) When SAP Group is moved to Node 2 successfully End users should automatically get connected to Node 2 without losing any data / locks.
Does our MSCS installtion is correct?
Thank you for your help.
Hello,
Yes in 640 system users get disconnected
note 112266
640 has an ABAP CI which is clustered . CI includes work processes as well as message and enque proc
With 700 situation is better
regards,
John Feely
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Question :-
1) Suppose , Both my SAP and SQL resources are running on Node 1. when I move SAP GROUP from Node 1 to Node 2, when it happens all end users get disconnected. Does it supposed to disconnect all end users when I move SAP GROUP from one node to other Node 2 ?
I thought failover mechanism should be transperent to users i.e when failover happens following things should happen
1)End Users should get "hour glass" .
2)All lock are preserved (enqueue)
3) When SAP Group is moved to Node 2 successfully End users should automatically get connected to Node 2 without losing any data / locks.
Answer :- Valid for NW04S (WAS 700 / 701):- The key is SMLG. You have to create Log on group in SMLG. Use this logon group in SAPLogon Pad. It behaves exactly the same way as I mention in my question i.e whenever failover occurs from Node 1 to Node 2 following things happened.
1)End Users should get "hour glass" .
2)All lock are preserved (enqueue)
3) When SAP Group is moved to Node 2 successfully End users should automatically get connected to Node 2 without losing any data / locks
While in 640 :- user get dissconnected when failover occurs
Hi,
Answer to 1) Yes, all users will be disconected, remember that central services are the resources moving but dispatcher stays in Node 1 so all users needs to reconnect.
The way the new switch over happens in Node 1 you are connected to a CI and when fail over you will be in a DI with a different dispatcher. Even Unix or Windows the users will be disconnected the virtual IP will fall and get up again. The Idea is that the downtime will be the minimal, if you do a fail over you will see a fast switch between nodes so the users will be able to work again in a minimal time.
Your installation is fine, the users will need to login again.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Mushtaq,
If you have CI in Node 1 and DI in Node 2 when the fail over happens then users will be disconnected. SAP person told us that the difference with the installation in MSCS in 6.40 is that in that version Message server, dispatcher and services moves all from Node 1 to Node 2 and it takes some minutes.
In 7.0 version message server and central services ASCS/SCS are located in shared disks, Central inscance and dispatcher, Dialog Instance and dispatcher are located in local disks so when a fail over takes place what moves are Central services ASCS/SCS and message servers so the fail over takes only 15 to 20 seconds.
If fail over happens then users will be disconnected because the message server gest disconnected for some seconds and gui gets disconnected.
If DB group fails over then mostly users won´t be disconnected.
Regards
User | Count |
---|---|
85 | |
10 | |
10 | |
9 | |
7 | |
6 | |
5 | |
5 | |
4 | |
3 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.