Forum Discussion
ictjl
Mar 02, 2018Altocumulus
Here is what I had to do to resolve the "shard" issue and get my DCD cluster back in good health. I had to rebuild the DCD cluster (only had 1 DCD), which means we lost data. However, I was in a pre-production so it's wasn't that big of a deal. I wasn't even able to remove the DCD from the centralized manager (CM). Kept getting a 503 error. The following curl command (removes indices) on the CM got me to where I could remove the DCD from the CM so I didn't have to rebuild the CM.
Note: You will lose data if you run this command
curl -X DELETE localhost:9200/*
Rebuild steps I took after running the curl command above on the CM:
- Remove the DCD from the CM (at this point I had 0 DCDs attached to my CM)
- Go to the DCD and reset to factory defaults (https://support.f5.com/csp/article/K15886) and go through the setup wizard (this kept the DCD master keys in tact so FPS on Big-IPs didn't need reconfigured)
- Upgrade both CM and DCD to latest version of Big-IQ (5.4 HF2)
- Attach DCD to CM and activate FPS service (FPS events started immediately processing)