Logan Attwood
Logan Attwood
https://github.com/lattwood/nomad/releases/tag/v1.7.8-yamux-fork I threw a build up there with the yamux fix.
Confirmed that hashicorp/yamux#127 fixes the issue. We'll be running a forked Nomad build until there's a release of Nomad with the yamux fix in it. edit: idk if I'm out...
@schmichael yup. (our nomad agents are distributed around the world)
Just saw another issue opened on yamux this week that could be responsible for what we're seeing here as well- https://github.com/hashicorp/yamux/issues/133
hmm, i don't think so? heh. On Thu, Sep 5, 2024 at 3:05 PM Michael Schurter ***@***.***> wrote: > Closed #23305 as > completed via hashicorp/yamux#127 > . > >...
In the context of this- https://github.com/hashicorp/nomad/issues/23848 would it be worthwhile for us to try dropping in the libp2p yamux fork
@schmichael I think the hung Read is a symptom, not the cause. I think the cause is a timed out write that the keepalive interval on the underlying session. https://github.com/hashicorp/yamux/issues/142
Fix opened- hashicorp/yamux#143
current diff I've compiled into a nomad binary and will be deploying shortly- https://github.com/lattwood/net-rpc-msgpackrpc/compare/master...lattwood:net-rpc-msgpackrpc:close_on_server_write_error
Keeping everyone updated- I've cut a custom binary and deployed it, branching from 1.9.3 with the fix.