Description
Describe the bug
Starting with one of the latest 2 or 3 releases, the IPFS companion seems to be disabled from time to time, even if Kubo is running.
When that happens, if I open a site such as ipfs.tech
, it'll load through HTTP instead of redirecting to http://ipfs.tech.ipns.localhost:8081
. If I hit the icon on the browser to open the companion menu, it'll then connect to Kubo and everything works as usual.
To Reproduce
I haven't found a pattern, it just happens sometimes... I can be sure that Kubo is running, however -- notice the Active line, it's been up for 2 weeks and 3 days:
$ systemctl --user status ipfs
● ipfs.service - InterPlanetary File System (IPFS) daemon
Loaded: loaded (/home/$USER/.config/systemd/user/ipfs.service; enabled; preset: enabled)
Active: active (running) since Sun 2023-09-10 20:13:23 WEST; 2 weeks 3 days ago
Docs: https://docs.ipfs.io/
Main PID: 1277 (ipfs)
Tasks: 139 (limit: 38327)
Memory: 652.9M (swap max: 0B)
CPU: 1h 59min 29.095s
CGroup: /user.slice/user-1001.slice/[email protected]/app.slice/ipfs.service
└─1277 /home/$USER/.local/bin/ipfs daemon --migrate
Sep 10 20:54:19 $HOSTNAME ipfs[1277]: ⚠️ Your system might be struggling to keep up with DHT reprovides!
Sep 10 20:54:19 $HOSTNAME ipfs[1277]: This means your content could partially or completely inaccessible on the network.
Sep 10 20:54:19 $HOSTNAME ipfs[1277]: We observed that you recently provided 129 keys at an average rate of 16.295281359s per key.
Sep 10 20:54:19 $HOSTNAME ipfs[1277]: 🕑 An attempt to estimate your blockstore size timed out after 5 minutes,
Sep 10 20:54:19 $HOSTNAME ipfs[1277]: implying your blockstore might be exceedingly large. Assuming a considerable
Sep 10 20:54:19 $HOSTNAME ipfs[1277]: size of 10TiB, it would take 379707h34m35.70358272s to provide the complete set.
Sep 10 20:54:19 $HOSTNAME ipfs[1277]: ⏰ The total provide time needs to stay under your reprovide interval (22h0m0s) to prevent falling behind!
Sep 10 20:54:19 $HOSTNAME ipfs[1277]: 💡 Consider enabling the Accelerated DHT to enhance your system performance. See:
Sep 10 20:54:19 $HOSTNAME ipfs[1277]: https://github.com/ipfs/kubo/blob/master/docs/config.md#routingaccelerateddhtclient
Sep 10 22:15:09 $HOSTNAME ipfs[1277]: 2023/09/10 22:15:09 http: superfluous response.WriteHeader call from github.com/prometheus/client_golang/prometheus/promhttp.(*responseWriterDelegator).WriteHeader (delegator.go:65)
(Those logs are from more than 2 weeks ago, they don't seem related? Could be wrong... I just didn't get around to taking care of that yet.)
Expected behavior
If Kubo is running, the companion should automatically connect and stay connected.
Desktop (please complete the following information):
- OS: Debian testing
- Browser: Firefox 119.0b2
- IPFS companion: 3.1.0