00:59:36 <barfield> bamahat: IDK if you have done any of the work already to replace stud in the portal, but I updated the portal-orchestration code to replace stud with haproxy. But I no longer have access to the repo to commit my changes. If you'd like the changes let me know. 
16:36:23 <barfield> I've got a triton compute node PI 20221215T000744Z for some reason when I rebooted into the newer version the VM hosted on that box stopped being accessible on the network. I determined there were too underlay vnics 0 and 1 and two different IP addresses at that. I do not think it was like that before the update. 
16:36:59 <barfield> There is only one underlay nictag however
16:38:14 <barfield> I deleted one of the underlays from napi, removed the VM's fabric nic, rebooted the compute node, and it went back to only underlay0. Then I added a new NIC to the VM. But it refuses to work. I am at a loss
16:40:20 <barfield> When I run dladm show-overlay I see the VM nic setup correctly, `vmadm get VM | json nics` looks correct, but you cannot ping the VM. I went in under single user mode in the guest and enabled networking (ubuntu 2204), the nic looks right, but I cannot ping anything on the fabric within the guest. 
16:40:49 <barfield> I can ping other hosts from the GZ on the underlay VLAN so it can't be an MTU/VLAN issue
18:11:45 <barfield> okay I found the problem with the overlay in portolan. Now to figure out how to fix it
18:51:22 <barfield> anyone have any idea how to reassociate an underlay nic with a compute node uuid in portolan?
18:51:50 <barfield> Everything looks to match other working compute nodes in sdc-napi.
19:38:15 <barfield> for future reference: portolan add-underlay --cn <UUID> --ip <cn_underlay_ip>
19:38:22 <barfield> boom back in business
19:42:43 <gonzosysadm[m]1> nice
19:43:33 <barfield> odd that portolan's path doesn't include the correct node path though
19:43:52 <barfield> I slapped a symlink in /opt/smartdc/portolan to /opt/smartdc/portolan/build/node
19:44:09 <barfield> Otherwise, the "portolan" command died with /usr/bin/env no such node
19:44:47 <barfield> Too bad theres no "portolan list-underlay"
19:44:59 <barfield> or help lmao
19:45:07 <bahamat> barfield: No idea what you're talking about there. If portolan didn't have the right path to node then the service would be completely broken.
19:45:22 <barfield> I think that smf calls the correct path
19:45:29 <barfield> but bash login doesn't
19:45:38 <barfield> well
19:45:41 <barfield> let me rephrase that
19:45:42 <bahamat> It's not supposed to.
19:45:54 <barfield> /opt/smartdc/portolan/node just doesn't exist
19:46:08 <barfield> but I found it in /opt/smartdc/portolan/build
19:46:22 <barfield> so cd /opt/smartdc/portolan; ln -s build/node
19:46:33 <barfield> makes `portolan` and `portolanadm` work
19:47:14 <bahamat> Ah.
19:47:35 <barfield> here is the PATH
19:47:36 <barfield> /usr/local/sbin:/usr/local/bin:/opt/local/sbin:/opt/local/bin:/usr/sbin:/usr/bin:/sbin:/opt/smartdc/portolan/node/bin:/opt/smartdc/portolan/node_modules/.bin:/opt/smartdc/portolan/bin
19:48:31 <barfield> Good thing I've spent some time playing with sdcnode recently lol
19:49:55 <barfield> and here it is:
19:49:56 <barfield> /usr/bin/ctrun -l child -o noorphan /opt/smartdc/portolan/build/node/bin/node --abort_on_uncaught_exception /opt/smartdc/portolan/server.js &
19:50:07 <barfield> /opt/smartdc/portolan/smf/method/portolan is what starts the SMF
19:53:26 <bahamat> Created TRITON-2359
19:53:43 <barfield> that was fast!
20:11:23 <bahamat> sdc-portolan#9
20:13:09 <bahamat> TritonDataCenter/sdc-portolan#9
20:13:09 <jinni> https://github.com/TritonDataCenter/sdc-portolan/issues/9