DataStore no longer knows its own name after Docker Engine was installed

146
2
Jump to solution
2 weeks ago
Brian_Wilson
Occasional Contributor III

I am running 11.2 DataStore on a machine "cc-gislicense" as Standby. The server is Windows Server 2019.

I had Docker Engine running  for a test, and I have removed it completely.

When it was installed, DataStore decided to change its own name from the FQDN of the server to HOST.DOCKER.INTERNAL. This is frustrating because now the primary datastore thinks that is the standby's name too, and it no longer resolves anywhere at all.

Some things I have tried include (reboot after each change)

I removed the entries that docker added in etc/hosts.

I put an entry in etc/hosts matching the IP address with the FQDN

I put "host_identifier=" with the FQDN in C:\Program Files\ArcGIS\DataStore\Framework\etc\hostidentifier.properties

Some change or another of the above caused logging and postgres to straighten themselves up but "describedatastore" persists in showing this

 

Information for relational data store ds_atjwr8b3
==================================================
Backup location.....................C:\arcgisdatastore\backup\relational
Is backup folder shared.............false
Backup schedule.....................{"schedule-starttime":"00:00:00","schedule-frequency":"Every 4 DAYS"}
Days backup retained................7
Data store status...................Starting
SSL enabled.........................true
Last failover.......................20240218174726754
member machines.....................HOST.DOCKER.INTERNAL CC-GISDATASTORE.CLATSOP.CO.CLATSOP.OR.US

 

 

0 Kudos
1 Solution

Accepted Solutions
Brian_Wilson
Occasional Contributor III

Since it's a backup machine anyway, I used the removemachine command, first with the real name, then with its assumed name.

removemachine cc-gislicense --store relational --force
removemachine HOST.DOCKER.INTERNAL --store relational --force

 

(Neither command worked without the --force option.)

Then running describedatastore on the primary no longer showed the useless name.

I can probably run configuredatastore to get the standby running again but for now I am going to leave it alone.

View solution in original post

0 Kudos
2 Replies
Brian_Wilson
Occasional Contributor III

My thinking is that it's so easy to rename a Datastore that it can happen with no operator intervention, so there must be a way to trigger it to make it go back to normal?

I guess it's also pretty easy to uninstall and reinstall Datastore since this is on a standby machine so should I try that?

 

0 Kudos
Brian_Wilson
Occasional Contributor III

Since it's a backup machine anyway, I used the removemachine command, first with the real name, then with its assumed name.

removemachine cc-gislicense --store relational --force
removemachine HOST.DOCKER.INTERNAL --store relational --force

 

(Neither command worked without the --force option.)

Then running describedatastore on the primary no longer showed the useless name.

I can probably run configuredatastore to get the standby running again but for now I am going to leave it alone.

0 Kudos