Home » Server Options » RAC & Failsafe » RAC - (Listener) - VIP doesn't migrate from one node to another (Oracle 10g 10.2.0.4 on AIX 6.1)
icon5.gif  RAC - (Listener) - VIP doesn't migrate from one node to another [message #551356] Tue, 17 April 2012 10:37 Go to next message
carlino70
Messages: 4
Registered: April 2012
Junior Member

hello people, I have the following problem on a 10.2.0.4 Oracle on AIX 1.6

I have 2 RAC nodes as the node # 2 goes down, the VIP migrated correctly to the node # 1, but the listener # 1 do not see the change, and just resolve to the local ip.

The same does not apply in the reverse case, when it drops the node N 1, the listener attends to the fallen VIP node and migrated.

The configuration is correct and identical in both nodes, according to the documentation, but I can not find a reasonable answer.

IP node # 1.
scm1or01 192.68.60.11
scm1or01-VIP 192.68.60.13

IP node # 2
scm1or02 192.168.60.12
scm1or01-VIP 192.168.60.14

Some idea of ​​the failure?, How I can solve it?

thank you very much

Below I write as I detect the problem:

when scm1or02 is down:


scm1or01:/cots/oracle/app/oracle/product/10.2.0/db_1/network/admin> lsnrctl status

LSNRCTL for IBM/AIX RISC System/6000: Version 10.2.0.4.0 - Production on 11-APR-2012 19:56:46

Copyright © 1991, 2007, Oracle. All rights reserved.

Connecting to (ADDRESS=(PROTOCOL=tcp)(HOST=)(PORT=1521))
STATUS of the LISTENER
------------------------
Alias LISTENER_SCM1OR01
Version TNSLSNR for IBM/AIX RISC System/6000: Version 10.2.0.4.0 - Production
Start Date 11-APR-2012 16:50:19
Uptime 0 days 3 hr. 6 min. 26 sec
Trace Level off
Security ON: Local OS Authentication
SNMP ON
Listener Parameter File /cots/oracle/app/oracle/product/10.2.0/db_1/network/admin/c
Listener Log File /cots/oracle/app/oracle/product/10.2.0/db_1/network/log/listener_scm1or01.log
Listening Endpoints Summary...
(DESCRIPTION=(ADDRESS=(PROTOCOL=tcp)(HOST=192.168. 60.13)(PORT=1521)))
(DESCRIPTION=(ADDRESS=(PROTOCOL=tcp)(HOST=192.168. 60.11)(PORT=1521)))
Services Summary...



scm1or01:/cots/oracle/app/oracle/product/10.2.0/db_1/network/admin> ifconfig -a
en7: flags=e080863,80<UP,BROADCAST,NOTRAILERS,RUNNING,S IMPLEX,MULTICAST,GROUPRT,64BIT,CHAIN>
inet 10.1.1.1 netmask 0xffffff00 broadcast 10.1.1.255
tcp_sendspace 131072 tcp_recvspace 65536 rfc1323 0
en6: flags=e080863,80<UP,BROADCAST,NOTRAILERS,RUNNING,S IMPLEX,MULTICAST,GROUPRT,64BIT,CHAIN>
inet 192.168.60.11 netmask 0xffffff00 broadcast 192.168.60.255
inet 192.168.60.13 netmask 0xffffff00 broadcast 192.168.60.255
inet 192.168.60.14 netmask 0xffffff00 broadcast 192.168.60.255
tcp_sendspace 131072 tcp_recvspace 65536 rfc1323 0
lo0: flags=e08084b,c0<UP,BROADCAST,LOOPBACK,RUNNING,SIM PLEX,MULTICAST,GROUPRT,64BIT,LARGESEND,CHAIN>
inet 127.0.0.1 netmask 0xff000000 broadcast 127.255.255.255
inet6 ::1%1/0
tcp_sendspace 131072 tcp_recvspace 131072 rfc1323 1

listener.ora #1 (idem listener.ora #2, except IPs)

LISTENER_scm1or01 =
(DESCRIPTION_LIST =
(DESCRIPTION =
(ADDRESS = (PROTOCOL = TCP)(HOST = scm1or01-vip)(PORT = 1521)(IP = FIRST))
(ADDRESS = (PROTOCOL = TCP)(HOST = scm1or01)(PORT = 1521)(IP = FIRST))
)
)

SID_LIST_LISTENER_scm1or01 =
(SID_LIST =
(SID_DESC =
(SID_NAME = xa21s)
(ORACLE_HOME = /cots/oracle/app/oracle/product/10.2.0/db_1)
(GLOBAL_DBNAME=xa21bcs_DGMGRL)
)
(SID_DESC =
(SID_NAME = xa211)
(ORACLE_HOME = /cots/oracle/app/oracle/product/10.2.0/db_1)
(GLOBAL_DBNAME=xa21_DGMGRL)
)
(SID_DESC =
(SID_NAME = xa212)
(ORACLE_HOME = /cots/oracle/app/oracle/product/10.2.0/db_1)
(GLOBAL_DBNAME=xa21_DGMGRL)
)
(SID_DESC =
(SID_NAME = xa211s)
(ORACLE_HOME = /cots/oracle/app/oracle/product/10.2.0/db_1)
(GLOBAL_DBNAME=xa21bcs_DGMGRL)
)
(SID_DESC =
(SID_NAME = xa212s)
(ORACLE_HOME = /cots/oracle/app/oracle/product/10.2.0/db_1)
(GLOBAL_DBNAME=xa21bcs_DGMGRL)
)
)

Also the listener.ora in scm1or01 starts how LISTENER, but the crsstat says:

scm1or01:/cots/oracle/app/oracle/product/10.2.0/db_1/network/admin>crsstat

listener_scm1oro1 ONLINE

I do not understand the reason, I've rebuilt the configuration of the listener, and netmgr netca tools, but the problem persists

Thanks in advanced!
Carlino
Re: RAC - (Listener) - VIP doesn't migrate from one node to another [message #551360 is a reply to message #551356] Tue, 17 April 2012 11:01 Go to previous messageGo to next message
John Watson
Messages: 4793
Registered: January 2010
Location: Global Village
Senior Member
Welcome to the forum. Please read our OraFAQ Forum Guide and How to use [code] tags and make your code easier to read

I don't see what the problem is: it looks like the normal behaviour, the VIP has failed over. Assuming that your clients have tnsnames entries that include both VIPs, with failover=on and load_balance=on, they will always get to a working listener. Can you post the output of
crs_stat -t
and the tnsnames.ora file? (Not forgetting that you MUST enclose them correctly with [code] tags, or they'll be really hard to read. Just as hard as the stuff you have already posted.)
Re: RAC - (Listener) - VIP doesn't migrate from one node to another [message #551385 is a reply to message #551360] Tue, 17 April 2012 14:52 Go to previous messageGo to next message
carlino70
Messages: 4
Registered: April 2012
Junior Member

Thank you for your answer.

VIP migration does not work alone in a sense, that is when the node scm1or01 falls. In this situation the VIP node should migrate down to be served by the node scm1or02.

While IP is OK with an "ifconfig-a", is never served by the active listener.

I copy here the tnsnames.ora

I understand that the processes Global Cache Services (GCS) and Global Enqueue Service (GES) should route the new configuration on the active node, but ... as I can check?

I hope you understand the problem, you could help me with some idea of ​​what happens.

thanks again!


# Tnsnames.ora for a RAC PCS system

xa21 =
  (DESCRIPTION =
	(ENABLE=BROKEN)
    (ADDRESS = (PROTOCOL = TCP)(HOST = scm1or01-vip)(PORT = 1521))
    (ADDRESS = (PROTOCOL = TCP)(HOST = scm1or02-vip)(PORT = 1521))
    (LOAD_BALANCE = no)
    (CONNECT_DATA =
      (SERVER = DEDICATED)
      (SERVICE_NAME = xa21pcs)
      (FAILOVER_MODE =
        (TYPE = SELECT)
        (METHOD = BASIC)
        (RETRIES = 5)
        (DELAY = 5)
      )
    )
  )

es =
  (DESCRIPTION =
    (ADDRESS = (PROTOCOL = TCP)(HOST = scm1or01-vip)(PORT = 1521))
    (ADDRESS = (PROTOCOL = TCP)(HOST = scm1or02-vip)(PORT = 1521))
    (LOAD_BALANCE = no)
    (CONNECT_DATA =
      (SERVER = DEDICATED)
      (SERVICE_NAME = xa21pcs)
    )
  )

isr =
  (DESCRIPTION =
	(ENABLE=BROKEN)
    (ADDRESS = (PROTOCOL = TCP)(HOST = scm1ed01)(PORT = 1521))
    (LOAD_BALANCE = no)
    (CONNECT_DATA =
      (SERVER = DEDICATED)
      (SERVICE_NAME = isrpcs)
    )
  )

####USED BY DBAs and SCRIPTS ONLY

xa211 =
  (DESCRIPTION =
    (ADDRESS = (PROTOCOL = TCP)(HOST = scm1or01-vip)(PORT = 1521))
    (CONNECT_DATA =
      (SERVER = DEDICATED)
      (SERVICE_NAME = xa21)
      (INSTANCE_NAME = xa211)
    )
  )

xa212 =
  (DESCRIPTION =
    (ADDRESS = (PROTOCOL = TCP)(HOST = scm1or02-vip)(PORT = 1521))
    (CONNECT_DATA =
      (SERVER = DEDICATED)
      (SERVICE_NAME = xa21)
      (INSTANCE_NAME = xa212)
    )
  )

###REMOTE LISTENER

listeners_xa21 =
  (ADDRESS_LIST =
    (ADDRESS = (PROTOCOL = TCP)(HOST = scm1or01-vip)(PORT = 1521))
    (ADDRESS = (PROTOCOL = TCP)(HOST = scm1or02-vip)(PORT = 1521))
  )

listeners_isr =
  (ADDRESS_LIST =
    (ADDRESS = (PROTOCOL = TCP)(HOST = scm1ed01)(PORT = 1521))
  )

Re: RAC - (Listener) - VIP doesn't migrate from one node to another [message #551390 is a reply to message #551385] Tue, 17 April 2012 15:32 Go to previous messageGo to next message
John Watson
Messages: 4793
Registered: January 2010
Location: Global Village
Senior Member
Quote:
I understand that the processes Global Cache Services (GCS) and Global Enqueue Service (GES) should route the new configuration on the active node
No, GCS and GES have nothing to do with networking. They are part of the RAC database instances; VIPs are part of the clusterware.

Apart from that, if I read your post correctly, I think you are expecting the listener to start listening on two VIPs following a VIP fail over. Wrong. The only purpose of VIP failover is so that if a node fails, sessions will detect the error immediately, and trigger TAF.

Lastly, your tnsnames entries could be better. You need load_balance=on to distribute the connection requests randomly. failover=on by default, but you should still specify it, to make the file self documenting.

One last point: you really should think about replacing your 10.2 clusterware with 11.2 (you don't have to upgrade the databases) 11.2 is lot more versatile and stable.
Re: RAC - (Listener) - VIP doesn't migrate from one node to another [message #551457 is a reply to message #551390] Wed, 18 April 2012 08:24 Go to previous messageGo to next message
carlino70
Messages: 4
Registered: April 2012
Junior Member

"Apart from that, if I read your post correctly, I think you are expecting the listener to start listening on two VIPs following a VIP fail over. Wrong. The only purpose of VIP failover is so that if a node fails, sessions will detect the error immediately, and trigger TAF."

But that if it works when the node scm1or02 falls over or fails then scm1or02-vip VIP migrates scm1or01 node.

I can verify this with:

lsnrctl status
LSNRCTL for Linux: Version 10.2.0.3.0 - Production on 18-APR-2012 13:14:32

Copyright (c) 1991, 2006, Oracle. All rights reserved.

Connecting to (DESCRIPTION = (ADDRESS = (PROTOCOL = TCP) (HOST = scm1or01) (PORT = 1521) (IP = FIRST)))
STATUS of the LISTENER
------------------------
Alias ​​LISTENER
TNSLSNR Version for Linux: Version 10.2.0.3.0 - Production
Start Date 09-APR-2012 21:14:24
Uptime 8 days 16 hr. 0 min. 9 sec
Trace Level off
Security ON: Local OS Authentication
SNMP OFF
Listener Parameter File / cots/oracle/app/oracle/product/10.2.0/db_1/network/admin/listener.ora
Listener Log File / cots/oracle/app/oracle/product/10.2.0/db_1/network/log/listener.log
Listening Endpoints Summary ...
  (DESCRIPTION = (ADDRESS = (PROTOCOL = tcp) (HOST = scm1or01) (PORT = 1521)))
  (DESCRIPTION = (ADDRESS = (PROTOCOL = tcp) (HOST = scm1or01-vip) (PORT = 1521)))
  (DESCRIPTION = (ADDRESS = (PROTOCOL = tcp) (HOST = scm1or02-vip) (PORT = 1521)))
Services Summary ...
Service "Xa21" has 1 instance (s).
  Instance "Xa21", status UNKNOWN, has 1 handler (s) for this service ...
Service "xa21bcs_DGMGRL" has 1 instance (s).
  Instance "xa21s", status UNKNOWN, has 1 handler (s) for this service ...
Service "xa21pcs_DGMGRL" has 2 instance (s).
  Instance "xa211", status UNKNOWN, has 1 handler (s) for this service ...
  Instance "xa212", status UNKNOWN, has 1 handler (s) for this service ...
The command completed successfully



ifconfig-a
eth0 Link encap: Ethernet HWaddr 00: A1: B0: D6: 1F: 8B
          inet addr: 192.168.60.11 Bcast: 192.168.60.255 Mask: 255.255.255.0
          inet addr: 192.168.60.13 Bcast: 192.168.60.255 Mask: 255.255.255.0
          inet addr: 192.168.60.14 Bcast: 192.168.60.255 Mask: 255.255.255.0
          inet6 addr: fe80 :: 2a1: b0ff: fed6: 1f8b/64 Scope: Link
          UP BROADCAST RUNNING MULTICAST MTU: 1500 Metric: 1
          RX packets: 878977 errors: 0 dropped: 0 overruns: 0 frame: 0
          TX packets: 382 errors: 0 dropped: 0 overruns: 0 carrier: 0
          collisions: 0 txqueuelen: 1000
          RX bytes: 91248265 (87.0 MiB) TX bytes: 41490 (40.5 KiB)
          Interrupt: 177 Base address: 0xde00



If you say is correct I should not see what I expect to see, in a sense of failure or the other, right?

Please clarify you can last, but is expected behavior should incorporate TAF in my configuration.

Thank you very much again, you're very kind!

Carlino

[Updated on: Wed, 18 April 2012 08:26]

Report message to a moderator

Re: RAC - (Listener) - VIP doesn't migrate from one node to another [message #551493 is a reply to message #551457] Wed, 18 April 2012 12:50 Go to previous message
carlino70
Messages: 4
Registered: April 2012
Junior Member

Here the crsstat command output:
[code]
scm1or01:/xahome/oracle> crs_stat -t
Name Type Target State Host
------------------------------------------------------------
ora....SM1.asm application ONLINE ONLINE scm1or01
ora....01.lsnr application ONLINE ONLINE scm1or01
ora....r01.gsd application ONLINE ONLINE scm1or01
ora....r01.ons application ONLINE ONLINE scm1or01
ora....r01.vip application ONLINE ONLINE scm1or01
ora....SM2.asm application ONLINE ONLINE scm1or02
ora....02.lsnr application ONLINE ONLINE scm1or02
ora....r02.gsd application ONLINE ONLINE scm1or02
ora....r02.ons application ONLINE ONLINE scm1or02
ora....r02.vip application ONLINE ONLINE scm1or02
ora.xa21.db application ONLINE ONLINE scm1or01
ora....11.inst application ONLINE ONLINE scm1or01
ora....12.inst application ONLINE ONLINE scm1or02
scm1or01:/xahome/oracle>


scm1or02:/xahome/oracle> crs_stat -t
Name Type Target State Host
------------------------------------------------------------
ora....SM1.asm application ONLINE ONLINE scm1or01
ora....01.lsnr application ONLINE ONLINE scm1or01
ora....r01.gsd application ONLINE ONLINE scm1or01
ora....r01.ons application ONLINE ONLINE scm1or01
ora....r01.vip application ONLINE ONLINE scm1or01
ora....SM2.asm application ONLINE ONLINE scm1or02
ora....02.lsnr application ONLINE ONLINE scm1or02
ora....r02.gsd application ONLINE ONLINE scm1or02
ora....r02.ons application ONLINE ONLINE scm1or02
ora....r02.vip application ONLINE ONLINE scm1or02
ora.xa21.db application ONLINE ONLINE scm1or01
ora....11.inst application ONLINE ONLINE scm1or01
ora....12.inst application ONLINE ONLINE scm1or02
scm1or02:/xahome/oracle>


[code]
Previous Topic: oracleasm scandisks issue
Next Topic: Voting disk backup
Goto Forum:
  


Current Time: Sun Nov 23 06:29:21 CST 2014

Total time taken to generate the page: 0.13839 seconds