pepoviola
commented
8 months ago Hi @Overkillus, thanks for the feedback and reporting this issue. I added a new config key (in settings) that allow you to disable the verification of the nodes. e.g
[settings]
timeout = 1000
node_verifier = "None"If is not set, the default behavior is the current one using prometheus server to query the startTime metric.
This is already available in the latest version v1.3.93.
Thx!
Issue Description
When simply spawning networks it stalls at
Error fetching metrics from: http://127.0.0.1:45667/metricsespecially for bigger networks. Sometimes it never completes and aborts network creation.It might be a natural consequence of starting a big network which just takes a while to boot up but then the error is not particularly telling of what's actually happening.
Steps to reproduce the issue
Steps to reproduce the issue
zombienet spawn <> -p nativeSometimes after printing errors for 30s+ it goes through the finishes up booting the network. Generally for bigger networks it seems to stall for longer often going over the limit and aborting.
Config used: https://github.com/paritytech/polkadot-sdk/blob/master/polkadot/zombienet_tests/functional/0002-parachains-disputes.toml
It is not unique to the above one and I have pretty much the same effect on all other network definitions of that size.
Describe the results you received
Below is the standard output of a
spawnbut at one point it simply stalls at:Error fetching metrics from: http://127.0.0.1:xxxxx/metricsrepeating it hundreds of times.Full output: https://pastebin.com/HageqCnD
Describe the results you expected
Network spinning up or at least understanding what is it getting stuck at.
Zombienet version
1.3.92
Provider
Native
Upstream Latest Release
Yes
Additional environment details
Ubuntu, Thinkpad X1
Additional information
Alice log:
Bob log:
Eve log:
Collator log:
All other instances of logs are pretty much identical to one of those 4.