action #62756
closed
openqa-clone-job downloads images very slowly
Added by dancermak over 4 years ago.
Updated about 4 years ago.
Description
When using openqa-clone-job
to clone a job from o3 to a local instance, I frequently have to download the Tumbleweed iso and hdd image. Unfortunately, openqa-clone-job
is very slow when downloading iso/qcow2 files, the download speed barely reaches 1 MB/s, while when downloading the same file via Firefox, my download speed reaches 10 MB/s.
If there is some throttling, could it be turned off?
There's no intentional throttling. Not sure why it is so slow in your case. I can usually get rates > MiB/s with openqa-clone-job
. Note that openqa-clone-job
uses internally LWP::UserAgent
to do the download so if there's something wrong with the download the problem is likely caused by that Perl module.
- Category set to Support
- Status changed from New to Feedback
- Assignee set to okurz
This could depend on multiple factors why it is slow for you. Is this to a machine within SUSE nbg network or elsewhere?
okurz wrote:
This could depend on multiple factors why it is slow for you. Is this to a machine within SUSE nbg network or elsewhere?
It is elsewhere, but geographically relatively close.
dancermak wrote:
okurz wrote:
This could depend on multiple factors why it is slow for you. Is this to a machine within SUSE nbg network or elsewhere?
It is elsewhere, but geographically relatively close.
yeah well, it must be "logically" close in the network to matter :) So, how else can we help you?
okurz wrote:
yeah well, it must be "logically" close in the network to matter :) So, how else can we help you?
Well, how can I debug the slowness of the download script? Because Firefox downloads at nearly full network speed.
try to create a small reproducer script, e.g. strip out the relevant part from the perl script. Or try with "curl" on the console if it is as fast as firefox or as slow as clone-job.
- Due date set to 2020-02-18
- Status changed from Feedback to Resolved
I've actually encountered this too on the staging machines a few months ago and it turned out to be a network problem then.
Also available in: Atom
PDF