on 02-22-2013 9:40 PM
Hi All,
We are building our SAP NW PI 7.31(ABAP+JAVA) as a HA system on sun-solaris 3.3. But we are facing issue in keeping services like ABAP Msg
Service(ms), ABAP enque service(en), Java-ms, Java-en in cluster.
Does oracle-solaris cluster 3.3 provide all the necessary cluster resources for NW 7.31 (Dual Stack) ? If yes could you please guide us to the relavant
procedure/documentation ?
We are working on :
OS : Solaris SPARC 10
DB : Oracle 11.2.0.2
Cluster : Oracle Solaris 3.3
SAP : NW PI 7.31 Dual Stack
Please advice...
Thanks & Regards,
Srikanth G
Hi,
What is the problem that you are facing in HA setup?
Thanks
Sunny
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Sunny,
Thanks for responding !!
From SAP note 1740958, understood that Oracle Solaris cluster 3.3 5/11 supports SAP Kernel 720_EXT.
We have two hosts running on cluster and installed ASCS, SCS, ERS(ASCS/SCS), DB, PAS on host1 and we are planning to run ERS (A)SCS services on host2 and rest on host1. Now we are trying to put each service in cluster where we had some confusion.. Agents provided by cluster 3.3 are SUNW.sapwebas, SUNW.sapscs, SUNW.sapenq and SUNW.repl, so could you please help us understand what are all cluster resources we need to create and which agent will take care of which service (i.e ASCS, SCS, ERS for ASCS, ERS for SCS, DB and PAS).
We have earlier followed a configuration guide which says to create separate resource for Orcl DB, Orcl lsnr, message server, enque server, ERS etc.., nothing mentioned about Java so we tried creating separate resource for Java ms, en and ERS but java ms was not coming up. For SAP Kernel 720_EXT I feel there would be some different procedure as we have ASCS/SCS scenario which clubs ms and en services.
Please advice..
Thanks & Regards,
Srikanth
Hi,
Please check below article:
http://docs.oracle.com/cd/E19787-01/820-7362/6nikish84/index.html
Thanks
Sunny
Hi Sunny,
We have created the below cluster resources one each for ASCS msg server, ASCS enq, ASCS ERS, SCS ms, SCS enq, SCS ERS, PAS(webas), DB, DB lsnr.
=== Cluster Resources ===
Resource Name Node Name State Status Message
------------- --------- ----- --------------
prd-webas-res nplspiapa2 Offline Offline
nplspiapp2 Offline Offline
prd-SCSmsg-res nplspiapa2 Offline Offline
nplspiapp2 Offline Offline
prd-SCSenq-res nplspiapa2 Offline Offline
nplspiapp2 Offline Offline
prd-ASCSmsg-res nplspiapa2 Offline Offline
nplspiapp2 Offline Offline
prd-ASCSenq-res nplspiapa2 Offline Offline
nplspiapp2 Offline Offline
prd-SAP-oracle-res nplspiapa2 Online Online
nplspiapp2 Offline Offline
prd-SAP-oralisten-res nplspiapa2 Online Online
nplspiapp2 Offline Offline
prd-SAP-hafoip2-res nplspiapa2 Online Online - LogicalHostname online.
nplspiapp2 Offline Offline - LogicalHostname offline.
prd-SAP-hastorageplus-res nplspiapa2 Online Online
nplspiapp2 Offline Offline
prd-SAP-hasp-rs nplspiapa2 Online Online
nplspiapp2 Online Online
prd-SCSrep-res nplspiapa2 Offline Offline
nplspiapp2 Offline Offline
prd-ASCSrep-res nplspiapa2 Offline Offline
nplspiapp2 Offline Offline
prd-SAP-rep-hafoip2-res nplspiapa2 Offline Offline - LogicalHostname offline.
nplspiapp2 Online Online - LogicalHostname online.
We were able to start the services in cluster but Java Server node is coming up.
dev_disp logs says :
"
Thu Feb 28 11:41:56 2013
***LOG Q0I=> NiPConnect: 10.142.1.132:3900: connect (146: Connection refused) [nixxi.cpp 2898]
*** ERROR => NiPConnect: SiConnect failed for hdl 48/sock 16
(SI_ECONN_REFUSE/146; I4; ST; 10.142.1.132:3900) [nixxi.cpp 2898]
*** ERROR => MsIAttachEx: NiBufConnect to ppidba-pub/3900 failed (rc=NIECONN_REFUSED) [msxxi.c 713]
"
dev_jstart log says :
"
M [Thr 01] Thu Feb 28 11:43:08 2013
M [Thr 01] ***LOG Q0I=> NiPConnect: 10.142.1.132:3201: connect (146: Connection refused) [nixxi.cpp 2898]
M [Thr 01] *** ERROR => NiPConnect: SiConnect failed for hdl 76/sock 19
(SI_ECONN_REFUSE/146; I4; ST; 10.142.1.132:3201) [nixxi.cpp 2898]
M [Thr 01] *** ERROR => EncNiConnect: unable to connect (NIECONN_REFUSED) [encomi.c 446] "
Dev_server0 log says :
"
F [Thr 17] Thu Feb 28 10:56:00 2013
F [Thr 17] *** LOG => SfCJavaVm: exit hook is called. (rc = -1501)
F
F ********************************************************************************
F *** ERROR => Java node 'server0' terminated with exit code -1501.
F ***
F *** Please see section 'Java program exit codes'
F *** in SAP Note 1316652 for additional information and trouble shooting advice.
F ********************************************************************************
F
F [Thr 17] *** LOG => exiting (exitcode -1501, retcode 2).
M [Thr 17] CCMS: CCMS Monitoring Cleanup finished successfully.
"
dev_server0.out says :
"
AS Java version [7.31.3710.148365.20111012091351] is starting server process [ID24102550] on host [ppidba-pub.northerngas.co.uk] ...
Loading [LogManager]... [310] ms
Loading [PoolManager]... [11] ms
Loading [ApplicationThreadManager]... 7.383: [GC 7.430: [ParNew: 343552K->4622K(381696K), 0.1752770 secs] 343558K->4629K(2059008K), 0.1979181 secs] [Times: user=0.14 sys=0.59, real=0.23 secs]
[1557] ms
Loading [ThreadManager]... [59] ms
Loading [ClassLoaderManager]... [65] ms
Loading [ClusterManager]... [926] ms
Loading [LockingManager]...
FATAL: Server process failed to start; manager [LockingManager] could not be started
Heap
par new generation reserved 419840K, committed 419840K, used 145795K [0xfffffffe80010000, 0xfffffffe99a10000, 0xfffffffe99a10000)
eden space 343552K, 41% used [0xfffffffe80010000, 0xfffffffe889ed318, 0xfffffffe94f90000)
from space 38144K, 12% used [0xfffffffe974d0000, 0xfffffffe97953950, 0xfffffffe99a10000)
to space 38144K, 0% used [0xfffffffe94f90000, 0xfffffffe94f90000, 0xfffffffe974d0000)
concurrent mark-sweep generation reserved 1677312K, committed 1677312K, used 6K [0xfffffffe99a10000, 0xffffffff00010000, 0xffffffff00010000)
concurrent-mark-sweep perm gen reserved 1048576K, committed 1048576K, used 17236K [0xffffffff00010000, 0xffffffff40010000, 0xffffffff40010000)
"
I have switched off the cluster and started the system manually but still the same error.
Is 3201 and 3900 the right ports system is looking out for ? Could you please advice what could be the issue ?
Instance number for ASCS is 00, SCS - 01, PAS - 02
Thanks in Advance !!!
Regards,
Srikanth G
The link provided by Sunny Pahuja contains all required information to cluster setup. You need to learn more about cluster configuration. In two words, for AS ABAP you need to create 3 resource groups (as minimum):
1) central resource group (ASCS) with resources of types: enqueue server(SUNW.sapenq), message server(SUNW.sapscs), logical host resource. Place it on node1.
2) replica resource group (ERS) with resources of types: replica server(SUNW.saprepl), logical host resource. Place it on node2.
Configure all required dependencies between them: they must run on different nodes, in case of failover of central resource group it moves to replica resource group, after that the replica resource group moves to another node (if exists).
3) optional: db resource group: place here db itself and listener. Place it on node 3. In failover case move it on node 4.
You doesn't need to place application servers in cluster (not SPOF).
All the same you can apply to java resources:
1) SCS
2) ERS
3) optional: DB
Regards,
Roman
Hi,
Could you please help us understand the importance of the ports 3201 and 3900 and how they would be configured as we could see netstat -an on these ports returns no value in fact none of ports are responding (/etc/services has an entry of 3201 port but not 3900). Do we need to make any settings so that these ports get recognized even from virtual IP ?
Secondly on HA setup, we have an ECC6.0 EHP5 system which was already running on HA and we are replicating a similar kind of setup for PI7.31 as well. So our Solaris Consultant created three resource groups as follows :
1. prd-SAP-rg (SUNW.sapscs:3.3, SUNW.sapenq:3.3, SUNW.sapwebas:6, SUNW.oracle_server:8, SUNW.oracle_listener:5, SUNW.HAStoragePlus:9),
2. prd-SAP-hasp-rg (SUNW.HAStoragePlus:9)
3. prd-SAP-rep-rg (SUNW.saprepl:4.1, SUNW.LogicalHostname:4)
And corresponding cluster resources created for each resource group are as below :
1. prd-SAP-rg
prd-webas-res
prd-SCSmsg-res
prd-SCSenq-res
prd-ASCSmsg-res
prd-ASCSenq-res
prd-SAP-oracle-res
prd-SAP-oralisten-res
prd-SAP-hafoip2-res
prd-SAP-hastorageplus-res
2. prd-SAP-hasp-rg
prd-SAP-hasp-rs
3. prd-SAP-rep-rg
prd-SCSrep-res
prd-ASCSrep-res
Kindly let us know if this is valid setup for HA.
Thanks in Advance !!!
Regards,
Srikanth
Hi Sunny,
Apologies for troubling you again... I have gone through these documents earlier and my only concern is PAS. I understand that PAS is no longer a SPOF and so it does not need to be part of the cluster. But our solaris consultant is keen to mime the setup which we have in our ECC environment for uniformity. So just need to understand if there would be any impact if Primary Application Server is also a part of the cluster..
Thanks in Advance !!
Regards,
Srikanth
User | Count |
---|---|
84 | |
24 | |
12 | |
9 | |
7 | |
6 | |
5 | |
5 | |
4 | |
4 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.