akash-network
[BUG] withdrawal TX occurs every 10m despite minimum-balance is back above the minimum-balance (10 AKT)
Yet to be confirmed but this is what I've been observing earlier today when setting mine to 24h instead of 24h0m0s.
Setting withdrawalperiod=24h via Helm (either in the values.yaml or on the cli via --set ..)
gets into the Pod as expected => OK
root@node1:~# kubectl -n akash-services exec -ti $(kubectl -n akash-services get pods -l app=akash-provider --output jsonpath='{.items[0].metadata.name}') -- env | grep -E 'AKASH_WITH'
AKASH_WITHDRAWAL_PERIOD=24h
However, what I've noticed is this:
AKASH_WITHDRAWAL_PERIOD=24h => withdrawal is happening every 10 minutes !
AKASH_WITHDRAWAL_PERIOD=24h0m0s => withdrawal is happening every 24 hours as expected
Got drained quit a few of AKT's since yesterday..
root@node1:~# echo "$(echo '"height","code","timestamp","tx","fee","type"'; kubectl -n akash-services logs $(kubectl -n akash-services get pods -l app=akash-provider --output jsonpath='{.items[0].metadata.name}') | grep 'txhash' | grep -Ewo '[0-9A-F]{64}\b' | uniq | while read TX; do akash query tx $TX -o json; done | jq -r '[ .height, .code, .timestamp, .txhash, (.tx | (.auth_info.fee.amount[0].amount|tonumber / pow(10; 6)), (.body.messages[] | ."@type")) ] | @csv')" | column -t -s ","
"height" "code" "timestamp" "tx" "fee" "type"
...
...
...
"5788554" 0 "2022-05-08T09:12:26Z" "84FAD9ED1EB18169338E77735A9B97781DA3F168CB966627F04CB6CDB2DF56F1" 0.003653 "/akash.market.v1beta2.MsgWithdrawLease"
"5788554" 0 "2022-05-08T09:12:26Z" "64B649D56B2CB6A5F0FC13B0B6FAE46A10BD232223EFB8DA9797D8CADC4D11F0" 0.003667 "/akash.market.v1beta2.MsgWithdrawLease"
"5788654" 0 "2022-05-08T09:22:28Z" "57C9175ACD66714262BD2E8E5BD0F2EA56F1C6E79F96198A67E1F5951950483F" 0.003653 "/akash.market.v1beta2.MsgWithdrawLease"
"5788654" 0 "2022-05-08T09:22:28Z" "7B887564D94916B1086F931AE443CE4F3439386AA7523F96338834D679394990" 0.003667 "/akash.market.v1beta2.MsgWithdrawLease"
"5788754" 0 "2022-05-08T09:32:29Z" "03357EA8119C85053FEE7F6F2F287BF4FE862174ED001E8CE3F05287C35CA171" 0.003653 "/akash.market.v1beta2.MsgWithdrawLease"
"5788754" 0 "2022-05-08T09:32:29Z" "81166BEF9C6406A5EF059B9484D518ED403022AC638EA0BCCF93224CEEB0CD5A" 0.003667 "/akash.market.v1beta2.MsgWithdrawLease"
so am now running helm-charts with --set withdrawalperiod=24h0m0s
just like suggested here:
$ akash provider run --help |grep withdr
--withdrawal-period duration period at which withdrawals are made from the escrow accounts (default 24h0m0s)
am not getting drained every 10 mins after that.
Have just launched it with 24h again and cannot see the same behavior.
I remember I've seen less than 10 AKT balance on my provider, so I think it has something to do with the minimum balance (10 AKT default) parameter -- most likely when the balance goes below that, it starts withdrawing the lease every 10 mins -- yet to be confirmed:
$ akash provider run --help |grep withdr
--minimum-balance uint minimum account balance at which withdrawal is started (default 10000000)
Balance is ~49 AKT, though:
I[2022-05-12|14:58:13.939] balance below target amount, withdrawing now module=provider-service cmp=balance-checker
I[2022-05-12|15:08:13.940] balance below target amount, withdrawing now module=provider-service cmp=balance-checker
I[2022-05-12|15:18:13.941] balance below target amount, withdrawing now module=provider-service cmp=balance-checker
I[2022-05-12|15:28:13.941] balance below target amount, withdrawing now module=provider-service cmp=balance-checker
Confirmed to be a bug here https://github.com/ovrclk/akash/blob/da9659ebba788974cb37e60f0ecba2f5056acde9/provider/balance_checker.go#L135 https://github.com/ovrclk/akash/pull/1576/files#r871518811
Greetings, seems I encountered same issue on docker version
balance below target amount, withdrawing now module=provider-service cmp=balance-checke
https://github.com/orgs/ovrclk/packages/container/akash/22445711?tag=0.16.4-rc2
may I know it is fixed on 0.16.4-rc2? or is my config issue..?
Tried both but seems still try to withdraw every 10mins when my provider is below 10atk (values.yml using in helm deploy)
#withdrawalperiod: "24h0m0s"
withdrawalperiod: "60m0s"
root@akash-provider-6bfcc8cb5b-d4zg5:/# env|grep -i with
AKASH_WITHDRAWAL_PERIOD=60m0s
root@akash-provider-6bfcc8cb5b-d4zg5:/# akash version
v0.16.4-rc2
