1. Assuming your backup is a 4.9 DB with dynamic roles (which at some point in time (when upgrading ACS from 4.x to 4.9) you have created by converting static roles to dynamic role via the script) - I guess it would be enough to rename/delete the following file: " */etc/cloudstack/management/commands.properties*". Read this for the full understanding please: https://www.shapeblue.com/dynamic-roles-in-cloudstack/
2. Not sure what can be the issue here, but assuming your agent.properties contains the "host" line pointing to the correct IP of your mgmt server (or the LB if you use that...) - they should reconnect just fine - after making sure that the mgmt server is working fine etc, restart agents on KVM hosts (you are using KVM?) and they should reconnect - if not, post here (actually on pastebin.org or similar) the agent logs (and mgmt logs) - so we can check. 3.For the mshost table, I would try stopping the mgmt and truncating the mshost table (make backup before...) and start mgmt from zero. Other way to remove mgmt from the mshost table would be to set the state to Down (or potentially something else...) and set the removed date - otherwise your new MGMT server will (based on the record in the mshost table) try to communicate with your old/dead mgmt server (with the same IP !!!) so this might be the problem obviously. Try to solve 1, then 3, and then 2 should be working on it's own possibly. Cheers On Wed, 11 Sep 2019 at 09:22, Ian Service <iserv...@ts2.ca> wrote: > Trying to clean up from a pooched hard drive, database backups are a-okay > and the cluster is running just fine but rebuilding my cloudstack > management server and I’ve got a couple errors I can’t get around. > > 1. I can’t log into the UI, the admin login works but then I get a 432 > error (The user is not allowed to request the API command or the API > command doesn’t exist) when asking for command=listCapabilites so it just > sits at the login screen. I looked into the role_permissions table and it > was empty, this system had been upgraded from 4.2 or so and now it’s at 4.9 > > 2. the management server doesn’t look like it’s able to connect with the > hosts, all operations are timing out. I’ve checked firewalls and routing, > same box, same IP, new drives > > 3. the mshost table has a new entry for the same host now, different msid > that doesn’t match the MAC address. Likely when adding this management > server back to the database it created that, how do I assume the old one? > > Any assistance would be appreciated, > > Thanks, > > - Ian > -- Andrija Panić
