have anyone gotten "there is no
have anyone gotten "there is no container instance that can be provided to this durable object" after deployed?
21 Replies
Was this immediately after first deploy?
We need to make this more obvious in UI, but when we first deploy, we are setting up/fetching containers in right spots
not exactly, my first deploy was 20 minutes ago
so there's a little delay between whent he Worker is live and when it can request a DO
hrm...
and the problem persists
for any Container?
max_instances count not set low?
yes, the instance is created but becomes instantly inactive
hrm...
Can you try @Cloudflare/containers 0.0.10?
Thats an earlier version*
THat should keep them alive properly
sorry bug that we're working on
I'm on 0.0.16 hehe
But... if they go invactive, it should just come back up
Yeah still try 0.0.10 sorry
(the real fix is still in flight - I need to update channel on this)
will try it rn
same problem @Mike Nomitch :/
Damn
ANything in the logs?
I wonder if its an OOM
Some context here: https://discord.com/channels/595317990191398933/1387061528016195684/1389320169851195423
FWIW this is one of the issue we're running into as well.
This happens with 5 instances available in the dashboard in an "Inactive" state on both cloudflare/containers 0.0.10 and 0.0.16.
0 logs from last 6 hours
well, the instance is being created as inactive and actually never becomes active
I think it doesnt even get time to use some memory
this is an example
same problem
is it the same bug that you guys are working on?
I need to know if I can rely on cloudflare containers or if should find alternatives
@Gabriel Francisco we're be digging in here, but still unclear on root cause. If its the OOMs, fix coming late this week/early next
If its something else, then TBD
(but given that 3 of you have hit it, we're not ignoring this)
If anybody here is able to try with
"instance_type": "standard" and let me know if that works, would be helpful for debugging
could help rule out OOMing and/or CPU issues
@ldl @Gabriel Francisco @Tory
We just realized that if the user doesn't set "max_instances", it get set to 0. - So it might be that. If you have it unset, try setting to 10 or something and seeing what happens.
If so: dumb bug 🤦♂️ but easy fixprob it, will try it now
Thanks