-
Notifications
You must be signed in to change notification settings - Fork 3
"Done downloading" but without any data #7
Comments
Hello, you're doing nothing wrong and you aren't alone with this problem. I'm investigating it. But I think it isn't container problem as it's only invoking ulozto-downloader with arguments. Seems like ulozto-downloader exits without any error message. But yes, there is one mistake in container. Script run.sh isn't checking return code so it shows I'll investigate it further and let you know. |
Found out two issues when using
Script needs to wait for flaresolverr to start. It worked in past because we weren't caching tordb. But now start of ulozto-downloader is too fast and flaresolverr isn't started yet.
|
Hi, thanks for reply, so is there some solution how to using container? thanks |
Yes. Working on it and will be commiting it in few hours. EDIT: Just pushing hotfix with sleep 10 seconds before starting ulozto-downloader. Should solve your issue and I'll try to make it more robust later. |
i tried rebuild it, but same issues, but it takes +10seconds longer, but no download starting, no downloaded data:
|
I could reproduce it myself. It happens to me only occasionally (1 of 50 tries) but error is clear: This error is not visible when ulozto-downloader is started with I need to figure out how to detect if Flaresolverr is started OK and listening... |
Added waiting for Flaresolverr start by checking its /health endpoint. I hope it will fix this issue. Please test and let me know. :) |
i tested it, it looks different but with same results:
|
I really don't know now. Everything is working on my Windows and Linux machines.
|
i tried to rebuild with --no-cache, i edit just run.sh, in others files there is no more edits:
|
Looks like that problem is with ulozto-downloader itself. Could you run container with
|
sure:
|
Yeah, that's the problem. Your CPU doesn't support AVX instructions needed for tensorflow, see #3 Because you have KVM you could try to emulate AVX with qemu. But I don't have any experience with that so you are on your own there. Another option is to compile tensorflow without AVX when building image. I tried in #no-avx branch but had no luck and interest in it. I'll welcome PR if you try to solve it. |
looks that you're right: i changed my CPU architecture and now it's works. if i run docker. Thanks so much for your patience
|
That looks promising. Now it works? |
yes, now it works <3
|
Sweet! 👍 |
HI, thanks for docker container. If i run it:
:~# time docker run -e PARTS=30 -e DEBUG=1 -e ENFORCE_TOR=1 -v /fastdata/Serialy:/downloads uld-docker:latest "https://ulozto.cz/file/TKvQVDFBEhtL/debian-9-6-0-amd64-netinst-iso"
Debug enabled!
Starting FlareSolverr
Starting ulozto-downloader for https://ulozto.cz/file/TKvQVDFBEhtL/debian-9-6-0-amd64-netinst-iso
09/24/23 17:50:46 - Done downloading
real 0m1.451s
user 0m0.014s
sys 0m0.023s
no more logs, no download data, container is in exited state.
what i do wrong?
thanks
The text was updated successfully, but these errors were encountered: