[tor-mirrors] A call for volunteers for mirror maintenance
sebastian at torproject.org
Fri Sep 9 11:22:25 UTC 2016
> On 08 Sep 2016, at 14:45, Sebastian Hahn <sebastian at torproject.org> wrote:
>> On 08 Sep 2016, at 02:47, Sebastian Hahn <sebastian at torproject.org> wrote:
>> as you've realized it's been taking forever to get mirror updates sent
>> to the website, sorry for that. If there's anyone who wanted to take
>> over, please drop me a line and I'll explain the process.
> Wow! I did not expect this kind of response, a total of 8 people
> have indicated interested so far, and the first batch of changes
> has already been applied. Thanks to all who wants to help, I
> appreciate it a lot.
> : https://trac.torproject.org/projects/tor/ticket/20108
Here's an email I sent to the first two people who mailed me. Hopefully
it explains the process a little.
both of you have expressed interest for maintaining the mirror list.
Thanks! I hope at least one of you will still want to do it after this
short introduction, if both of you do that's fine, too :)
The website is maintained in a git repo. The metadata about mirrors
is stored there, along with a perl script to sanity-check them and
maintain the list. The fiile include/tor-mirrors.csv contains the raw
data as submitted by the operators, include/mirrors-table.wmi gets
auto-generated from that file by using the update-mirrors.pl script.
When running the script, it fetches a sample of files from the official
website and then compares them to the files received from the mirrors.
If there are any discrepancies, these are reported/excluded. The entire
process is fragile and many improvements can be made, but it should
One issue that can happen is that if the perl LWP module cannot verify
the certificate of https://torproject.org, all mirrors are deleted from
the output file. In this case, make sure to set appropriate environment
variables (these depend on the LWP version) such that certificates can
be checked correctly.
What I would hope to get from you is that you take the input from people
on the mailing list, put them into the csv file, make a git commit, run
the script, check the result for errors, make another git commit, and
then submit a patch to the git repository by creating an issue on
trac or sending me email directly.
Let me know if you have any questions/concerns. Thanks again!
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 455 bytes
Desc: Message signed with OpenPGP using GPGMail
More information about the tor-mirrors