Disclamer: The following steps regarding the repointing of the VxRail Manager to a new PSC are not supported nor official and should be performed under supervision of trained professionals / certified engineers and/or in a training environment! These steps were performed in my own lab and should not be performed in a production infrastructure without contacting the support!
After successfully converging of the external PSC into the vCSA and decommissioning of the old external PSC, it is time to tell the VxRail Manager VM that things have changed.
I tried to install the update locally through the VxRail Manager and it looked good until it reached 65%. Then the process stops.
I opened two putty sessions to the VxRail Manager and checked for the “web.log” and “lcm.log”. The “lcm.log” showed, that it tried to contact the old PSC and failed.
2019-11-11T14:40:37.261+0200 INFO [localhost-startStop-1] com.vce.commons.core.services.ConnectionHelper ConnectionHelper.refreshConnection:34 – VC connection timed out, reconnecting.
2019-11-11T14:40:37.262+0200 INFO [localhost-startStop-1] com.vce.commons.core.services.ConnectionHelper ConnectionHelper.refreshConnection:36 – Connecting to vcsa.com.com username vxrailboss@vsphere.local
2019-11-11T14:40:37.268+0200 INFO [localhost-startStop-1] com.vce.commons.core.connection.misc.ConnectionUtils ConnectionUtils.getSsoUrl:124 – Trying to obtain sso token from: https://psc.com.com/sts/STSService/vsphere.local
2019-11-11T14:40:40.828+0200 ERROR [localhost-startStop-1] com.vce.commons.core.connection.services.ConnectionFactoryImpl ConnectionFactoryImpl.createVCConnection:106 – unable to connect to VC
com.sun.xml.ws.client.ClientTransportException: HTTP transport error: java.net.NoRouteToHostException: No route to host (Host unreachable)
So I checked the VxRail Manager again:
# psql -U postgres mysticmanager
# select * from settings;
# select * from management_account;
Note the id number in the row for the PSC. In my case it is id = 1. After that I checked the ‘runtime.properties’. I know it says “do not modify”, but since it is a test environment, I just testing. Backups were made beforehand.
Next I created a new password hash for the management user. In my case the ‘vxrailboss@vsphere.local’ using ‘echo -n “MyNewPassword” | openssl bf-ecb -nosalt -K $(xxd -p < /etc/vmware-marvin/password.key) | xxd -p’ and saved the output. using ‘vi /var/lib/vmware-marvin/runtime.properties’ I changed the ‘data.vcenter.pscHost=’ pointing to the vCSA running with the embedded PSC and the ‘data.managementPassword=blowfish\:’ to the newly created passhash.
The same password I just created the passhash for needs to be configured in the vCSA for this particular user.
Back to the SSH session, the information in the Postgres databases need to be updated also. So I checked my notes about the row id = 1, the new FQDN of the vCSA with the embedded PSC and the passhash created earlier.
# psql -U postgres mysticmanager
# update settings set psc_ip = ‘vcsa.com.com’;
# update management_account set host = ‘vcsa.com.com’ where id = 1;
# update management_account set password = ‘11111111111111111111’ where id = 1;
Reviewing the command showed all rows are now updated. Note that the password row changed to the new one and the host row is pointing to the ‘vcsa.com.com’.
Repointing is done, but the update was started earlier. So this process needs to be stopped. So again, I checked the Postgres database:
# psql -U postgres mysticmanager
# select id, components, upgrade_status, upgrade_filepath from virtual_appliance where component = ‘VXRAIL’;
And changed it with:
# psql -U postgres mysticmanager -c “update virtual_appliance set upgrade_status = ‘HAS_NEWER’ where component_id = ‘VXRAIL_SYSTEM’;”
After refreshing the VxRail Manager page the update process can be restarted and worked like a charm.
I again want to make clear that these steps are not official supported! Use these information on your own risk!
Personal Blog about vAdventures 