High CPU Usage when 'tailscaled' service is stopped

[sygibson]

I am running Trayscale v0.12.3 via flathub. TailscaleD version is 1.64.0.

When I stop tailscaled (systemctl stop tailscaled), my Trayscale service starts consuming between 10 to 12% of my CPU causing laptop fans to spin up to compensate for the heat.

If I start tailscaled back up - then it returns to normal CPU usage.

I believe this is a distinct issue from #66 reported issue.

Is there any more information/details I can provide to help?

[DeedleFake]

I've just run into this one myself. It's probably related to the recent Mullvad changes.

If you enable Tailscale from the Trayscale UI, does Trayscale freeze up on you completely? It does for me.

[DeedleFake]

This seems to be kind of inconsistent. I was unable to replicate it on another machine. Very strange.

[pavelskipenes]

My system just paniced and I had to do a forced reboot. Checking the latest logs I saw only messages from Trayscale.

ERROR get Tailscale status err="get tailscale status: Failed to connect to local Tailscale daemon for /localapi/v0/status; not running? Error: dial unix /var/run/tailscale/tailscaled.sock: connect: no such file or directory"

Out of curiosity I counted them and Trayscale produced 10 million log lines and occupied 99.915% of my journal file that boot 😅

journalctl -b -1 | grep dev.deedles.Trayscale.desktop | wc -l
Journal file /var/log/journal/8efa45f8c2da413ea33beb33a4b647b8/[email protected]~ is truncated, ignoring file.
10078907

journalctl -b -1 | wc -l
Journal file /var/log/journal/8efa45f8c2da413ea33beb33a4b647b8/[email protected]~ is truncated, ignoring file.
10087447

>>> total = 10087447
>>> trayscale = 10078907
>>> trayscale / total
0.9991534032347332

[DeedleFake]

Oh my. Are you starting it using a systemd unit? I use .config/autostart and it doesn't put any logs into journald.

You're not the first person to have logging issues. I'll have to take a look at reducing the amount of log output or producing a log file instead of outputting to stdout/stderr or something. Ironic considering that I've usually had issues with not enough logging in projects before...