Why does the Perspective URL change to an IP address on failover?
The browser opens the session with a hostname, for example https://IS1M:8043/data/perspective/client/MyProject. When the backup takes over, the address bar changes to the backup gateway's IP address. When the master returns, it changes to the master's IP instead of its hostname.
The cause is the gateway's public address setting. When it is left on Automatic, each gateway advertises the address it detects on its own interface, which is the IP. On failover, the running Perspective session is sent to whatever address the redundancy pair advertises for the active node. The browser follows that address, so the IP replaces the hostname you typed. To fix it, set the public address on each node manually to that node's own hostname, set the ports explicitly, and restart both gateways.
Which hops carry the session from master to backup?
Follow the packet. Four points in the path decide which address the browser lands on.
| Hop | What it contributes | Failure if misconfigured |
|---|---|---|
| Client DNS resolver | Resolves IS1M and IS2B to IPs |
Hostname redirect fails even when the gateways are set correctly |
| Master gateway public address | Address advertised for the master node | Failback lands on the master's IP |
| Backup gateway public address | Address advertised for the backup node | Failover lands on the backup's IP, or on the dead master |
| Backup redundancy connection to master | How the backup reaches the master to sync and monitor it | Inconsistent addressing between nodes; set it to the master's hostname |
Check layer one first. From the client machine, resolve both hostnames (nslookup IS1M, nslookup IS2B) and open each gateway's page by hostname on both ports. If either hostname fails from the client, no gateway setting will produce a hostname redirect. With the browser's developer tools open on the network tab, force a failover and read the redirect target. That shows directly whether the gateway is advertising an IP or a hostname.
Which configuration approach fits this redundancy pair?
| Approach | Running session after failover | New session while master is down | Result |
|---|---|---|---|
| Public address = Automatic (default) | Transfers, but the URL shows an IP | User must type the backup URL | The symptom described above |
| Manual; both nodes set to the master's hostname | Redirect target is the node that just failed | User must type the backup URL | Did not work in the tested pair |
| Manual; each node set to its own hostname, ports 8088/8043 | Transfers to IS2B by hostname and back to IS1M on failback |
User must type the backup URL | Confirmed working |
| Perspective Workstation configured with both gateway links | Transfers | Workstation holds both links, similar to the Vision Client Launcher | Best fit for operator stations |
| One browser URL that loads either gateway | Ignition has never provided this through the public address. It needs infrastructure outside Ignition (DNS or a proxy layer). | Outside gateway configuration | |
Recommendation: give each node its own hostname as its public address, with explicit HTTP and HTTPS ports. On fixed operator stations, add Perspective Workstation with both gateway links so a fresh launch works while the master is offline.
How do you set per-node public addresses?
- On the master, open the gateway settings and change the public address from Automatic to manual.
- Enter the master's hostname (
IS1Min this example). Set the HTTP port to8088and the HTTPS port to8043. A full URL typed into the host field while the ports stay wrong or automatic was not enough in the tested case. The ports had to be set explicitly. - On the backup, repeat the change using the backup's own hostname (
IS2B), not the master's. Use ports8088and8043. - In the backup's redundancy settings, point the connection to the master at the master's hostname, not its IP.
- Restart the master gateway, then the backup. The public address change may not take effect until after a restart.
- Confirm that every client subnet can resolve both hostnames before you run the failover test.
What recurs on redundant Perspective pairs?
| Pitfall | Symptom | Correction |
|---|---|---|
| Setting changed, no restart | Still redirects to the IP | Restart both gateways |
| Only the master updated | Failover shows the IP, failback shows the hostname | Configure the backup with its own hostname |
| Backup public address set to the master's hostname | Session is sent back toward the failed node | Backup advertises IS2B
|
| TLS certificate issued for hostnames only | Certificate warning after an IP redirect on 8043
|
Fix the public address. Include both hostnames in the certificates each node serves. |
| Expecting the bookmarked master URL to open the backup | New browser tab fails while the master is down | Use the backup URL or Workstation with both links. A running session transfers on its own; a new one does not. |
| Setting not found after an upgrade (reported on 8.1.47) | Public address seems to be missing | Find the setting in that version's gateway settings documentation. A missing field does not by itself cause the "same URL for both nodes" limitation. |
How do you verify failover and failback URLs?
- From a client, open
https://IS1M:8043/data/perspective/client/MyProjectby hostname and confirm the session loads with no certificate warning. - Stop Ignition on
IS1M. Watch the address bar. It must change tohttps://IS2B:8043/..., not to an IP, and the session must keep running. - With the master still down, open a new tab to the
IS2BURL and confirm it loads. On Workstation stations, relaunch and confirm the backup link connects. - Start Ignition on
IS1Mand wait for the pair to return to master-active. - Confirm the running session returns to
https://IS1M:8043/...by hostname. In the browser network tab, confirm the failover and failback redirect targets both show hostnames.
FAQ
What happens if both redundant gateways use the master's hostname as the public address?
On failover, the session is directed toward the master's hostname, which is the node that just went down. Set each node's public address to its own hostname: IS1M on the master and IS2B on the backup.
What happens if I open the master URL in a browser while the master is offline?
The page does not load. Ignition has no public address feature that serves either gateway from one URL. Open the backup URL, or use Perspective Workstation configured with both gateway links.
What happens if I change the public address but the URL still switches to an IP?
Restart the gateway, because the setting may not apply until then. Also check that the ports are set explicitly to 8088/8043 and that the backup node was changed too.
Does a running Perspective browser session fail over automatically?
Yes. A running session transfers to the active node without user action. The public address only controls whether the redirect target shows the hostname or the IP.