Cluster failover scripts for LiveCache 7.4

Cluster failover scripts for LiveCache 7.4

Post by Martin Brod » Sun, 09 Mar 2003 02:21:06



Does anyone have experience with liveCache 7.4 failover scripts on
Tru64 Truclusters, if so how to implement it.

I am advised there are three files around named
dbserv.scr
lccluster
set_db_local_scr

does anyone have them or got them from some education/white papers
from SAP/Compaq/HP

Thanks

 
 
 

Cluster failover scripts for LiveCache 7.4

Post by Stephen Ker » Thu, 01 May 2003 22:27:36


Martin
I didn't use these scripts, as HP/Compaq wouldn't give them to us
without extortionate Consultation fees.

In the end, I used the standard cluster failover scripts.

You need one to stop LC and x_server, another to start x_server and LC.

Then another to perform the checking (as used by caad) of processes.
I set PROBE_PROCS to PROBE_PROCS="/sapdb/programs/pgm/vserver
/sapdb/LCT/db/pgm/kernel LCT"

However, you will also need to look at your lcinit script, at the
call_user_exit routine which references LCCLUSTER
(LCCLUSTER="$INSTROOT/sap/lccluster").

We have created our owne shell script lccluster.  It calls
caa_stop/caa_start as user <lc sid>adm.
You will have to do some Unix DOP (Division of Priveleges) config to
ensure that sidadm can run caa_stop/caa_start progs.

--
Posted via http://dbforums.com

 
 
 

1. Cluster failover without Cluster

Hello from France ...

After looking at cluster solutions on Solaris (Sun, Legato and Veritas),
we are a little disappointed because none of them can manage a HA level
(i mean 99.99... %), especially if storage is non-Sun and at a very
high cost (100.000 to 150.000 $ for the solution).

So has anybody implement a cluster-failover solution between
2 (or more) Sun Exxxx servers without any commercial cluster solutions.

I think that we can do a lot of things with Veritas VM using mirroring
between the 2 disk arrays (A3500FC or A5000). Of course, it becomes more
complex with the failover process between the 2 servers (IP, hostid,
apps
binaries, etc...).

Another question : is this the good way to have 2 equal servers to make
a good cluster (i mean not have a E10K in primary with a E3500 in
failover).

Thanks a lot for any answer.

2. Gnome Problems & X color depth

3. Sequent Cluster Failover Scripts

4. Help setting up RH7 on multiple drives

5. LPRng-3.7.4 and Mandrake-8.0??

6. Please help--Dates in shell scripts

7. TCP Wrappers 7.4

8. Q: 2.4 and Iomega Jaz drive

9. screen-3.7.4

10. eg fault with taper6.7.4

11. FreeBSD 7.4 i386 with PAE kernel crash

12. Q: taper-6.7.4 child segfault

13. TSM3.7.4 Netscapeand AIX4.3.3