Open andy108369 opened 1 year ago
it is still located https://github.com/akash-network/provider/blob/76acbe7736f8495fb8dd8f20b5c075395f42f455/cmd/provider-services/cmd/run.go#L610
and both env variables AKASH_CLUSTER_WAIT_READY_DURATION
AP_CLUSTER_WAIT_READY_DURATION
or flagcluster-wait-ready-duration
should work
it is still located https://github.com/akash-network/provider/blob/76acbe7736f8495fb8dd8f20b5c075395f42f455/cmd/provider-services/cmd/run.go#L610 and both env variables
AKASH_CLUSTER_WAIT_READY_DURATION
AP_CLUSTER_WAIT_READY_DURATION
or flagcluster-wait-ready-duration
should work
Yeah, they look like leftovers to me... Are they really used?
yes, tho people should use what is default value
yes, tho people should use what is default value
I agree. It makes me wonder why it was increased at all... Steven probably did that back then.
I'll drop it in the helm-charts provider so it gets to the default value.
There is a flag
AKASH_CLUSTER_WAIT_READY_DURATION
which we are setting to30s
in helm-chart based provider:Trying to figure out the difference between Praetor-based and helm-chart based provider install, I've only noticed
AKASH_CLUSTER_WAIT_READY_DURATION
is not being set on the former. As I've suggested Praetor team to try setting this parameter to60s
to rule out the issue where the leases would get removed upon provider (systemd service in their case) restart, I looked into the code and haven't found any reference this flag is used...The last reference I found is the one which was removed back in Akash
v0.16.5-rc0
according to this commit: https://github.com/akash-network/node/commit/f161334fec87a0c4b072077a9147d6cd21138427#diff-d9bf78faae6e633425de7ee74e6d6f7a1ffffebdb2e4e62bbef7f759d5783594L84-L93 (look forcfg.ClusterWaitReadyDuration
inprovider/service.go
)Do you think it should be put back? If not, then this flag should also be gone.