2 months ago
Hi, I’m trying to convert my Postgres instance to HA, but it failed with an error.
From the network tab:
```
{
"data": {
"workflowStatus": {
"status": "Error",
"error": "An unknown error occurred"
}
}
}
```
Background
Initially, I attempted the conversion but discarded the changes because renaming the cluster (from "Postgres HA" to "database") didn't seem to work.
After discarding, I can no longer attempt the HA conversion again.
Has anyone encountered this issue or knows how to reset/retry the conversion?
Attachments
6 Replies
Status changed to Awaiting Railway Response Railway • about 2 months ago
2 months ago
We weren't able to exactly reproduce this, but we believe we have merged a fix. Could you please try again?
Status changed to Awaiting User Response Railway • about 2 months ago
brody
We weren't able to exactly reproduce this, but we believe we have merged a fix. Could you please try again?
2 months ago
Still not working.
Would it be possible for you to take a look at my service and see if there’s anything misconfigured or causing the issue?
Here is the link to my service:
Status changed to Awaiting Railway Response Railway • about 2 months ago
Status changed to Awaiting User Response Railway • about 2 months ago
brody
Would we be able to try to initialize the HA Postgres ourselves?
2 months ago
Sure. Since this is a preview environment, I can redeploy the Postgres instance (non-HA) after you identify the root cause.
Status changed to Awaiting Railway Response Railway • about 2 months ago
2 months ago
Our engineer was able to upgrade the database to HA just fine, we believe you had a state frontend cached.
Status changed to Awaiting User Response Railway • about 2 months ago
brody
Our engineer was able to upgrade the database to HA just fine, we believe you had a state frontend cached.
2 months ago
Thanks! I checked and was able to roll back to a non-HA instance.
Just a suggestion: would it be possible to keep applications pointing to the original Postgres instance during the HA conversion process for a safer migration? Then users could switch environment variables after everything is fully set up.
Otherwise, it seems like some downtime is unavoidable during the transition (although changes can technically be discarded manually).
Status changed to Awaiting Railway Response Railway • about 2 months ago
Status changed to Solved injung • about 2 months ago