Hello, community! Happy new year! 🙇 We’re using AWS EKS for deploying datahub and got a problem with deleting metadata (glossary, domain..) I deleted all the glossary terms and domains with

--hard

option. When I check mysql database (we are using aws rds), all rows has been deleted. But! In UI, I could still see Glossary 1.8K and when I click it, there’s no glossary data. Same thing happens when I search for a word.. 😢 How could I clean up all these metadata? (I couldn’t try

datahub docker nuke

, since we are using k8s) Please help.. 🙏 Thanks in advance 🙇

try running the reindexing job https://datahubproject.io/docs/how/restore-indices/#kubernetes?

Thank you for replying! 🙇 We’ve tried reindexing command and also cleanup command. I could see those jobs(completed) with

kubectl get pods

command. But sadly it didn’t work 😢 Even if all the data is deleted, I want to delete them all.. 😢

NAME                                                 READY   STATUS      RESTARTS   AGE
...
datahub-cleanup-job-tn8k7                            0/1     Completed   0          4h25m
datahub-restore-indices-adhoc-nrnvm                  0/1     Completed   0          4h25m
....

Is it possible to go to ES and take a look at the glossary index to see if the docs are still inside? And the usual qn: which version are you on?

☝️ + Were your GMS pod restarted since then? Might be cached?

ohhhhhh!!! It’s solved!! 🙇 🙏 (Our datahub version was v0.9.3. I also tried restarting GMS pod with the command

kubectl rollout restart deployment datahub-datahub-gms

and it didn’t work.. 😢) I tried to search how to clean ES as xL said, and restoring indices cronjob with

-a clean

argument saved me!!! 🎉

-a clean

is written right above(in docker-compose), but I couldn’t think of applying it to kubernetes.. 🥲 Thank you so much for quick and amazing solution, xL and Peter 🙇 teamwork thanks bear

glad it worked out teamwork