I’m running a warp-cli client on my homelab server. I found warp-svc memory leak very fast.
3 Likes
Maybe I should use wgcf now 
I’m seeing this as well, and looking into having systemd kill it if it goes over 1GB or so. I’m using mode:proxy fwiw.
I’ve got a setup where cloudflared is running on a main machine and then I have WARP deployed to several clients to connect in to the main server. I’ve been having this issue for the past week and I’ve just been blaming my applications, but then I ran htop really fast and saw the WARP CLI was to blame.
I don’t know what’s going on, but the WARP CLI seems far from production ready if it’s having memory leakage issues this bad, and for it to not be fixed this long. I know I can’t say much because I’m on Cloudflare’s free plan, but it seems a bit silly that Cloudflare is making this stuff seem like production software when it isn’t even working correctly.
My only solution right now is similar to @tickets method, I’m just going to automatically be restarting the systemd service every hour so the memory can get freed up. I’m probably going to end up find some other service in the meantime though, it’s caused 20+ outages on my server this week (including on some things I have other people using) and it doesn’t appear Cloudflare has any plans on fixing this any time soon.