Issue with SCAN listener relocating

[Saud]

hi there,
i have 2 node RAc on 11 g R2 with Solaris 10 and ASM
3 of the SCan Listener are aval to Node1 and i want that 1 x SCAN listener should also be registered with NOde2 Instance. having said i have two question
1. even Node1 have all 3 SCAN listener registere to the instance but i still see sessions On NOde2, why ?
2. sir secondly when i try to relocate one SCAN listener to Node2 so am successfull but recive an error.

./crsctl status res -t and it shows that All endpoints are not registered  with SCAN_Listener1. i have check the google and it says that this is because the reason that u have an entry in the listener.ora which have the same ip address etc. so they suggesst that ist down the local Listener and then SCAN listener from GRID_HOME adnd then restart it from GRIID_HOME but this didnt work

please sir guide me

Thank you for your time

Regards

[ErmanArslansOracleBlog]

1. even Node1 have all 3 SCAN listener registere to the instance but i still see sessions On NOde2, why ?

- ANSWER: That's expected. That's because scan listeners redirect the connections to local listeners.. So the local listeners of both node 1 and node2 are registered with your scan listeners running on node 1 and those scan listeners redirect to connections to the nodes (to the local listeners on nodes) according to the load.

2. sir secondly when i try to relocate one SCAN listener to Node2 so am successfull but recive an error.

-ANSWER: Probably you have a misconfiguration in the hosts file or ip addresses that scan listeners are configured to use. (SCAN VIPs)

*Anything changed recently? Is this a new installation? Does this issue start to appear after a failover or reboot? Anyting changed with the RAC ip configurations?

In order to investigate this, we need the following from both of the nodes->

*crsctl status res -t
*lsnrctl status <SCAN_LISTENER_NAME> (for all the scan listeners)
*crsctl stat res -p |grep ENDPOINTS
*Scan listener logs.
*srvctl config scan
*ifconfig outputs

IMPORTANT NOTE : Don't upload log files and outputs that contain the hostnames or ip addresses the servers of your company. Mask your logs and outputs before sending them to me.
I would like to underline that I accept no responsibility about this. The risk is yours..

[Saud]

Sorry sir i was away for couple of weeks so couldn't answer.

Yes, there was major activity held is which we switch over to Huawei Storage from netapp and this also involves reboots of machineand rebalancing of Data so this issue is raised then

we are not connected to internet so i need to type the commnd reply manually or will take snap of the out put and attatch it.

[Saud]

crsctl status res -t

it has got big putput so i will write those which have the issue

ora.LISTENER.lsnr                   ONLINE   ONLINE            NODE1
                                             ONLINE   INTERMDIATE   NODE2

ora.LISTENER_SCAN1.lsnr       ONLINE   ONLINE            NODE1
ora.LISTENER_SCAN2.lsnr       ONLINE   ONLINE            NODE1
ora.LISTENER_SCAN3.lsnr       ONLINE   ONLINE            NODE1


rest everything is fine for this command..

2. snrctl status LISTENER_SCAN1 or2 or 3  it give foloowing messages
TNS-01101 : Could not find service name.
i have the path /u01/app/11.2.0.4/grid/admin here the scan listener listener.ora exist and all the 3 scan listener.


3.crsctl stat res -p |grep ENDPOINTS

ENDPOINTS=TCP:1521
ENDPOINTS=TCP:1528
ENDPOINTS=TCP:1521
ENDPOINTS=TCP:1521
ENDPOINTS=TCP:1521



4.

[Saud]

the out put of listener_scan1 is huge but writing the last line

<msg time='2020=02-04T22:37:47.559+05:00 'org_id=oracle' comp_id='tnslsnr' type='UNKNOWN'
level='16' host_id='xxxxxx' host_addr='x.x.x.x'>
<txt>04_Jan-2020 22:37:47 * (CONNECT_DATA=(CID=(PROGRAM=) (HOST=XXXX) (USER=XXXX)
(COMMAND= STATUS)(ARGUMENTS=64)(SERVICE=LISTENER_SCAN1) (VERSION=186647552))  * status * 0
</txt>
</msg>

[Saud]

6. srvctl config scan_listener
SCAN:Listener LISTENER_SCAN1 exists. Port: TCP:15xx
SCAN:Listener LISTENER_SCAN2 exists. Port: TCP:15xx
SCAN:Listener LISTENER_SCAN3 exists. Port: TCP:15xx



ifconfig -a
i have seen that all the 3 scan ips exists only on node1 by looking at the ifconfig -a and same are not available on 2nd node





please reply

[ErmanArslansOracleBlog]

I can't see the problem in full detail but there is bug record for this issue ->

Bug 15922206  All the SCAN VIPs start on the same node
Ref: Bug 15922206 - All the SCAN VIPs start on the same node (Doc ID 15922206.8)

Check the above document and see if it is similar to your issue.

As a workaround, you may relocate 1 scan vip using srvctl manually.

But this is a risky move, and the risk is yours!

Usage: srvctl relocate scan -i <ordinal_number> [-n <node_name>]
-i <scan_ordinal_number> Ordinal number of IP address
-n <node_name> SCAN VIP node name
-h Print usage

First check -> srvctl status scan
                     srvctl status scan

Example command ->  srvctl relocate scan -i 2 -n node2

Then check again->
                     srvctl status scan
                     srvctl status scan_listener

[Saud]

sir i have already did this but then what happened is that .let say like  if i shift one scan to node 2 so the remaining 2 x scan shows error in front of it which is "All end points are not registered". this is what i see infront of the 2 x scan listeners on node1 and on node 2 it is ok .

sir please give me a detaled document which is in sequence for understanding this local listener and reote listner in RAC as i see mnay documnets on web but still confuding.this database is running from last 10 yeasrs.

plz give me an article on lisetners in RAC so i can have deep underatnding of it as i am really confused.

Regards

[ErmanArslansOracleBlog]

Interresting..

As for getting info you can read this -> https://ermanarslan.blogspot.com/2014/02/rac-listener-configuration-in-oracle.html

Also this one -> NOTE:887522.1 - Grid Infrastructure Single Client Access Name (SCAN) Explained

In order to investigate this issue , we need all the related logs and traces.
We need to review TFA and/or RDA diags to get the whole picture..

Anyways, your issue is strange.. It seems to be caused by a structural misconfiguration.
As a quick check -> What is written in /var/log/messages of node 1?
Anything you can find in ->  CRS_HOME/log directory?

Info about RAC logs -> http://db.geeksinsight.com/2012/09/29/11gr2-rac-important-trace-directories-log-locations/