Context Deadline fails w/ statesync + arm

[dreamcodez]

Summary of Bug

Context deadline failure which impacts only arm with statesync

5:41PM INF error from light block request from primary, removing... error="post failed: Post \"https://rpc.alley.umeemania-1.network.umee.cc:443\": context deadline exceeded" height=831700 module=light primary={}
5:41PM ERR error on light block request from witness, removing... error="post failed: Post \"https://rpc.cope.umeemania-1.network.umee.cc:443\": context deadline exceeded" module=light primary={}
5:41PM ERR failed to remove witnesses err="no witnesses connected. please reset light client" module=light witnessesToRemove=[0]
Error: failed to start state sync: failed to set up light client state provider: post failed: Post "https://rpc.cope.umeemania-1.network.umee.cc:443": context deadline exceeded

I tested this with linux/amd64 via docker emulation and it does not have this problem.

Version

Please provide the output of the following commands:

• $ umeed version => 2.0.3-rc2
• $ go version => N/A (release build)
• $ uname -a => Darwin 21.4.0 Darwin Kernel Version 21.4.0: Fri Mar 18 00:46:32 PDT 2022; root:xnu-8020.101.4~15/RELEASE_ARM64_T6000 arm64

Steps to Reproduce

Steps to reproduce the behavior:

1. use an arm64 platform build
2. run umeed start w/ statesync
3. example genesis: https://raw.githubusercontent.com/umee-network/testnets/main/networks/umeemania-1/genesis.json
4. example config.toml:

[p2p]
persistent_peers = "[email protected]:26656,[email protected]:26656"
[statesync]
rpc_servers = "[https://rpc.alley.umeemania-1.network.umee.cc:443](https://rpc.alley.umeemania-1.network.umee.cc/),[https://rpc.cope.umeemania-1.network.umee.cc:443](https://rpc.cope.umeemania-1.network.umee.cc/)"
trust_height = 831700
trust_hash = "546155C86814F3DBFA31F0DBB6626C4160A73668763C7F8A43E8FCC1B01FD84A"
trust_period = "3000h"

---

For Admin Use

• Not duplicate issue
• Appropriate labels applied
• Appropriate contributors tagged
• Contributor assigned/self-assigned

[brianosaurus]

This is an HTTP-RPC timeout.

[dreamcodez]

This is an HTTP-RPC timeout.

I agree. However:

1. on arm it consistently fails after the timeout, every time -- no matter how many times i run
2. umeed status seems to also have a similar problem on arm, it hangs when trying to contact rpc
3. when i run emulated amd64 the problem dissappears

[brianosaurus]

You just need to increase the rpc_write_timeout in app.toml. I got the same error in an Intel i7 until I increased that. Alternatively, it could be the rpc_read_timeout. I haven't figured out which one exactly is the problem but it occurs on a POST so I figure write timeout is the culprit. It's also set to 0 seconds by default.