Open paulcostinean opened 5 months ago
thanks for reporting. Such a weird bug.
# Shard blocks across multiple store gateway instances. This option needs be set
# both on the store-gateway and querier when running in microservices mode.
# CLI flag: -store-gateway.sharding-enabled
[sharding_enabled: <boolean> | default = false]
I think this CLI should be -store-gateway.sharding-enabled=true
Short form works for toggling BoolVar
I just tried it and it is reproducible with the following:
- -store-gateway.sharding-enabled=true
- -store-gateway.sharding-ring.consul.hostname=<redacted>:8500
- -store-gateway.sharding-ring.replication-factor=3
Unfortunately I cannot reproduce this issue on my end. And I don't see how it matters by changing the order of the CLI arguments.
Describe the bug
Upgraded to 1.17 in one setup with no issues, but the querier started crashlopping in a second instance. The order of parameters seems to have some side effects.
To Reproduce Steps to reproduce the behavior:
Start Cortex (1.17) querier with the following arguments:
Perform Operations(Read/Write/Others) The querier returns the following error:
Expected behavior The querier should've started and
store-gateway
addresses should be fetched from the sharding ring (consul based)Flipping
store-gateway
specific arguments around worksFrom (broken):
To (working):
I think the struct used for
sharding-ring
is not initialised when setting thestore-gateway.sharding-enabled
parameter in the querier.Environment: