Skip to content

Pyramid v1.1.1 using all CPU resources and filling up memory #24

@FabioBonfiglio

Description

@FabioBonfiglio

My pyramid 1.1.1 instance, approx. 10 minutes after startup, starts to use 100% of all CPU cores, and progressively fills all the available memory.

Here are the logs, after 15 minutes:

Mar 21 20:55:04 neuchstr systemd[1]: Started pyramid.service - pyramid relay.
Mar 21 20:55:04 neuchstr pyramid-exe[16320]: 8:55PM INF running pyramid version=v1.1.1
Mar 21 20:55:04 neuchstr pyramid-exe[16320]: 8:55PM DBG calculated free ranges from index scan 10015555=1209 10165498=192 10173298=1363 10187258=952 10193748=278 10266381=249 10268983=383 10299166=1449 10302649=393 1039512=300 10403275=145 10407470=348 1>
Mar 21 20:55:04 neuchstr pyramid-exe[16320]: 8:55PM INF running on https://0.0.0.0:443 and http://0.0.0.0:80
Mar 21 20:55:04 neuchstr pyramid-exe[16320]: 8:55PM INF fetched latest version source= version=v1.1.1
Mar 21 20:57:04 neuchstr pyramid-exe[16320]: 8:57PM INF fetching primary follow lists for members n=27 relay=inbox
Mar 21 20:57:14 neuchstr pyramid-exe[16320]: 8:57PM INF fetching secondary follow lists for follows n=2125 relay=inbox
Mar 21 20:58:04 neuchstr pyramid-exe[16320]: 8:58PM INF checking ots proofs
Mar 21 20:58:04 neuchstr pyramid-exe[16320]: 8:58PM INF upgraded pending OTS proofs errored=0 pending=4 upgraded=0
Mar 21 21:01:27 neuchstr pyramid-exe[16320]: 2026/03/21 21:01:27 NOTICE from wss://relay.gathr.gives: 'error: internal error processing subscription 2236:repl~3'
Mar 21 21:04:50 neuchstr pyramid-exe[16320]: [khatru-relay] 2026/03/21 21:04:50 failed to upgrade websocket: write tcp 10.240.1.3:443->162.158.217.38:9967: write: connection reset by peer
Mar 21 21:04:50 neuchstr pyramid-exe[16320]: [khatru-relay] 2026/03/21 21:04:50 failed to upgrade websocket: write tcp 10.240.1.3:443->162.158.217.24:9506: write: connection reset by peer
Mar 21 21:04:51 neuchstr pyramid-exe[16320]: [khatru-relay] 2026/03/21 21:04:51 failed to upgrade websocket: write tcp 10.240.1.3:443->172.71.141.75:13003: write: connection reset by peer
Mar 21 21:08:16 neuchstr pyramid-exe[16320]: [khatru-relay] 2026/03/21 21:08:16 error writing ping: write tcp 10.240.1.3:443->23.129.64.152:19937: write: broken pipe; closing websocket
Mar 21 21:09:03 neuchstr pyramid-exe[16320]: [khatru-relay] 2026/03/21 21:09:03 error writing ping: write tcp 10.240.1.3:443->45.84.107.17:2461: write: connection reset by peer; closing websocket
Mar 21 21:09:37 neuchstr pyramid-exe[16320]: [khatru-relay] 2026/03/21 21:09:37 error writing ping: write tcp 10.240.1.3:443->185.220.101.35:10879: write: connection reset by peer; closing websocket

At this point, already 3.5 GB of RAM is used, and growing at a pace of ~10 MB per second.

To be able to restart pyramid, I always need to manually delete the mmmm.lock file.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions