yeah, ive done that but i keep getting
yeah, ive done that but i keep getting
Error: internal error
37 Replies
What's the id? Could it be some character in the id itself?
the durable object ID of the affected instance?
Yes. Since you are using Agent, I'm assuming you are using a human readable name rather than an opaque GUID
the ID is
smol-brains-10985
. the DO id is 0d98a1818d13e963b905dc11f22eddd1e51869ab8243bb9401b8085469576833
if that helps.
appreciate the help.I should have said, "name". Is that smol-brains-10985
yep
Is there any data in that instance?
cant check, because trying to access RPC methods just gives me internal server error
could check that instance's databaseSize in the GraphQL API
Can you push an update with an onRequest handler and hit that instead of using RPC?
That may not work because partyserver/set-name might actually be involved even with onRequest, but just giving you some more ideas to try
is this to just check if the request passes to onRequest or not?
something like this will suffice?
Sure
But if that works, then you can use that to inspect the data
Upgrade onRequest with code to inspect the data
Actually, I don't think you need to call super.onRequest
onRequest is a callback so the data comes through the Agent class's fetch handler so it gets a chance to do whatever it needs to do. This is why I don't think you should call super.onRequest
oh ok. what should i return here then?
return new Response("Hello from Agent!");
or something like thatgot it
I'm not familiar with CharacterAgent. I just use fetch to get to my Agent's onRequest callback
ok, still getting
this is the code:
its not going to the onRequest at all
it seems like the DO for this ID is just broken
What about
await characterStub.fetch(reconstructedRequest),
will try that out
Sorry, I cut and paste without editing. It's changed now. It now says,
.fetch(...
instead of .onRequest(...
yep
but failed again:
doesnt log
I'm out of ideas for now
gotta be borked
can a cf employee examine the issue using the reference id?
other "name"s log properly
cc @lambrospetrou
Firstly, don't tag individual CF people.
Secondly, I assume all the URLs you posted above are just fake/placeholders right, since they are under the main cloudflare domain?
Yes, placeholders
I shared the error with the team to see why it happens.
thank you, its urgent.
this is the reference ID if it helps, (error) Error: internal error; reference = 51a83596v7mruuigsoqssufi
i can share more details on DMs if it helps
The underlying root cause is the same as the other errors above. Can you describe a bit the workload? How many queries, how often, etc?
well, I can't even access this durable object ID's fetch handler
Did you create any virtual tables (FTS5) or any other schema on the actor's SQLite storage?
Any insight into the workload will help the team find the root cause faster.
no virtual tables. we are using SQLite schema migrations from your https://github.com/lambrospetrou/durable-utils
GitHub
GitHub - lambrospetrou/durable-utils: Utilities for Cloudflare Dura...
Utilities for Cloudflare Durable Objects and Workers - lambrospetrou/durable-utils
OK, thanks for the info.
fyi, it happened again on a different DO instance
@jcheese are you still experiencing the same DO internal error? I started getting them yesterday a few minutes after 11a EDT. I have an hourly cron trigger that invokes a websocket. Every time it fires, I get this message:
internal error; reference = kig34utl9ep9b0h2snkjpteo
The reference value is different each hour. Guessing that's simply a GUID for the internal error.
U doing any migrations? That was the issue for me
If by migrations you mean including this in my wrangler.toml, then yes:
durable_objects.bindings = [
{ name = "WEBSOCKET_HIBERNATION_SERVER", class_name = "WebSocketHibernationServer" }
]
migrations = [
{ tag = "V1_WEBSOCKET_HIBERNATION_SERVER", new_sqlite_classes = ["WebSocketHibernationServer"] }
]
How did you solve @jcheese ?no, i mean sql migrations directly inside the DO