-
Notifications
You must be signed in to change notification settings - Fork 1.7k
Parity can't handle too much RPC calls #6575
Comments
yeah, same issue happen when upgrade to v1.7.2 |
Some report: |
Are you re-using the same IPC connection or spawning a new one for each request? |
I think |
You should get much better performance by re-using connections, I can easily do 20k req/s using That said, we probably have a bug somewhere in IPC causing old connections to stay open. I think I saw another issue hanging around regarding this. |
This is the simplest IPC requester in python:
The result:
Parity:
If I use:
Result:
Something is wrong :( |
This one? #6516 |
I have recoded the |
Greetings.
Parity/v1.7.2-beta-9f47909-20170918/x86_64-linux-gnu/rustc1.19.0
Linux Debian Jessie.
Installed from dpkg binary.
Run with:
parity --no-ws --no-ui --tracing on --chain classic --base-path "/mnt/etc" --peers 50 --no-dapps
While scanning from 0 block to the current with
trace_block
, after 2-3k block i got this error and the clients crashing:I know this is a lot of RPC call pro second . I would like to catch every ether transfer to my db.
Update:
I use IPC to connect to the node.
Update 2:
When I put a 0.1 sec sleep between RPC calls then happens also..
The text was updated successfully, but these errors were encountered: