Hello everyone,
I deployed DataHub in GKE using He...
# troubleshootg
gentle-camera-33498
08/08/2022, 3:13 PMHello everyone,
I deployed DataHub in GKE using Helm. I have 4 replicas of GMS and 2 replicas of frontend, but I still have a lot of these errors. Can Anyone know what it could be?
gentle-camera-33498
08/08/2022, 3:21 PMI enabled cache on the search service and increased the batch return to 300. Can this change cause this problem?
o
orange-night-91387
08/08/2022, 3:36 PMDo you have the actual error occurring on GMS that is responding with a 500?
orange-night-91387
08/08/2022, 4:05 PMFrom this it looks like GMS is unable to connect to your database.
g
gentle-camera-33498
08/08/2022, 4:06 PMI understand. I'll try to update the cluster's secrets and see if that solves the problem.
gentle-camera-33498
08/08/2022, 4:07 PMBut wait.. The job of restoring the indexes is running successfully and uses the same access parameters as the GMS
gentle-camera-33498
08/08/2022, 4:10 PMeven doing the restoration of the indices, I constantly get this type of error:
o
orange-night-91387
08/08/2022, 4:21 PMA 500 error coming back from GMS will always have an underlying application error happening (barring 50X series that are specifically at the connectivity layer like gateway timeouts). There should be an error happening on GMS each time you get that, verify which one it is and that should tell you what's going on.
If you're consistently getting the connection down error, there is something going on with GMS's connection to your database. It doesn't make a lot of sense though if restore indices is working as they use the exact same configuration classes.
g
gentle-camera-33498
08/08/2022, 4:30 PMI'll try to see if it could be some connection/session configuration in the Cloud SQL instance that is hosting the MySQL database.
Thanks a lot @orange-night-91387
It has already given me a light where I can start trying to solve the problem teamwork