Details
-
Improvement
-
Resolution: Fixed
-
Major
-
2.7.0
-
Security Level: Public
-
CBG Sprint 42, CBG Sprint 43
-
3
Description
Testing CBS 6.5.0 w/ SGW 2.7.0 in a docker-compose environment, there is a 'gap' in the retry logic that SGW uses when connecting to a server. There is a retry loop where it tries to connect to the server and there is a 'Verifying index availability for bucket' phase which tests to see if the indicies it needs are defined and built. Between these phases, however, there is a period when the SGW can successfully connect to the server, but the bucket isn't ready yet, and so its attempts to run the test queries fail with
MCResponse status=KEY_ENOENT, opcode=0x89, opaque=0, msg: from [127.0.0.1:9101] |
It would help if SGW could anticipate this part of the potentially slow startup of the server. Log attached.