Mapping new Address - Fails in my setup

msanjib

Hi Team,
This is a nightmare for now to troubleshoot this issue. I have almost spent lot of time to figure out the reason for the failure. Any urgent help will be really appreciate, I am running out of time for a deadline.
I wanted to change the existing ip adresses of my setup ( network: 10.189.199/22 )to 3 new ip addresses (10.176.45.0/23).
I wanted use the re_ip with admintool. Generally it works but this time I am getting issue and almost no clue.
[Additional Info : These 3 vms are replicated from the primary site to disaster recovery site. Actually the idea, is that we want to use VM Ware storage level replication in active passive mode.]

Logs from adminTool.log

2020-08-09 13:37:03.618 admintools/18909:0x7f8b4b4ef740 [NodeState.transition] v_dummy_node0001=Preparing -> v_dummy_node0001=Prepared
2020-08-09 13:37:03.618 admintools/18909:0x7f8b4b4ef740 [NodeState.transition] v_dummy_node0002=Preparing -> v_dummy_node0002=Prepared
2020-08-09 13:37:03.618 admintools/18909:0x7f8b4b4ef740 [NodeState.transition] v_dummy_node0003=Preparing -> v_dummy_node0003=Prepared
2020-08-09 13:37:03.618 admintools/18909:0x7f8b4b4ef740 [NodeState.transition] v_dummy_node0002=Prepared -> v_dummy_node0002=OutOfDate
2020-08-09 13:37:03.618 admintools/18909:0x7f8b4b4ef740 [NodeState.transition] v_dummy_node0003=Prepared -> v_dummy_node0003=OutOfDate
2020-08-09 13:37:03.618 admintools/18909:0x7f8b4b4ef740 [NodeState.transition] v_dummy_node0001=Prepared -> v_dummy_node0001=RollingBack
2020-08-09 13:37:03.618 admintools/18909:0x7f8b4b4ef740 [NodeState.transition] v_dummy_node0002=OutOfDate -> v_dummy_node0002=RollingBack

2020-08-09 13:37:03.618 admintools/18909:0x7f8b4b4ef740 [NodeState.transition] v_dummy_node0003=OutOfDate -> v_dummy_node0003=RollingBack

As per my knowledge, the ssh connection goes fine, it read the input file , parse and then connect also. The issue is while updating the configuration file.
Please advise if there is any work around? I am just a learning Vertica and may be not knowing the internal of its working, however I will be interested.

I will also attach the full log if required.
Thanks in advance.
Regards,
SM

Hibiki

How did you shut down the cluster before changing the IP addresses? The status of node #2 and #3 was OutOfDate. It seems node #1 has the latest catalog and it hasn't been synced with the rest of nodes.

Jim_Knicely

I think the key reason why node's 2 and 3 are probably out of date is:

[Additional Info : These 3 vms are replicated from the primary site to disaster recovery site. Actually the idea, is that we want to use VM Ware storage level replication in active passive mode.]

Per the Vertica doc (https://www.vertica.com/docs/10.0.x/HTML/Content/Authoring/SupportedPlatforms/Virtualization.htm):

Hibiki

Hi Jim,
Thank you for your comment. Yes, that is why I wanted to know how to shut down the cluster.