Update .env for new server and agent image versions; enhance docker-compose.yml with dynamic agent names and extra hosts; add troubleshooting guide for port allocation issues.
This commit is contained in:
32
TROUBLESHOOTING.md
Normal file
32
TROUBLESHOOTING.md
Normal file
@@ -0,0 +1,32 @@
|
||||
# Troubleshooting
|
||||
|
||||
## Port 55432 Already Allocated (E2E Builds)
|
||||
|
||||
**Symptom:**
|
||||
A build step like "Postgres Up" fails with:
|
||||
|
||||
```
|
||||
Bind for 0.0.0.0:55432 failed: port is already allocated
|
||||
```
|
||||
|
||||
**Cause:**
|
||||
A Postgres container from a previous build was not cleaned up properly (e.g. build failed before reaching its teardown step). Since all E2E builds use the same hardcoded host port `55432`, the stale container blocks any new build from starting.
|
||||
|
||||
**Diagnosis:**
|
||||
|
||||
```bash
|
||||
# Find what's holding the port
|
||||
docker ps --filter "publish=55432" --format "table {{.Names}}\t{{.Status}}\t{{.Ports}}"
|
||||
```
|
||||
|
||||
**Fix:**
|
||||
|
||||
```bash
|
||||
# Remove the stale container (replace the name with the one from the command above)
|
||||
docker rm -f <container-name>
|
||||
```
|
||||
|
||||
Then re-run the failed build.
|
||||
|
||||
**Prevention:**
|
||||
Ensure every build configuration runs `docker compose down -v` in a cleanup/always-run step so containers are torn down even when earlier steps fail.
|
||||
Reference in New Issue
Block a user