[Pulp-list] pulp_streamer - The requested URL returned error: 400 Bad Request

[Matthew Madey]

Also, I noticed that if I create a repo without any feed, then use the
update command to add the feed\cert information later, it does not create
an additional pulp/repos/ directory under
/var/lib/pulp/published/yum/https/repos
when I synchronize the repository. I'm not sure if this is a bug but
something to be aware of if you are using another Pulp repo as the feed on
another Pulp server.

On Fri, Nov 2, 2018 at 12:36 PM Matthew Madey <mattmadey at gmail.com> wrote:

> This has been solved.. I was missing some options in the squid.conf. I
> copy/pasted the full config example from the documentation, restarted
> squid, and things started working as expected.
>
> On Tue, Oct 30, 2018 at 1:40 PM Matthew Madey <mattmadey at gmail.com> wrote:
>
>> I've been trying to get the deferred download features working on Pulp.
>> Currently I have a "master" Pulp server with a custom repo, then a "slave"
>> Pulp server that points to the "master" for the same content. I've
>> installed\configured squid\pulp_streamer on the "slave" and setup it's
>> custom repo to use on_demand download policy. Unfortunately, when a client
>> requests a package, I get the error: [Errno 14] PYCURL ERROR 22 - "The
>> requested URL returned error: 400 Bad Request"
>>
>> Some interesting things that I noticed..
>>
>> When setting up the custom repo on the "slave", I noticed that it created
>> an additional /pulp/repos/ directory
>> under /var/lib/pulp/published/yum/https/repos
>> Example:
>>
>> [root at slave repos]# pwd
>> /var/lib/pulp/published/yum/https/repos/pulp/repos
>>
>> This caused me to have to change the rewrite rule in the pulp_rpm.conf to
>> the below, otherwise when I would go to
>> http://slave.mycompany.com/pulp/repos/ I would have a pulp directory.
>> Clicking on that would go to a repos subdir, which had a pulp subdir.. on
>> and on.. because the rewrite rule would keep going back.. I'm thinking that
>> might be part of the issue I'm seeing. Is this expected behavior when you
>> synchronize one pulp repository with another pulp server?? I went down this
>> route as per documentation since the Parent\Child setup is deprecated and
>> very buggy..
>>
>> <Location /pulp/repos/>
>>   RewriteEngine On
>>   RewriteCond %{HTTPS} on
>>   RewriteRule (.+/pulp/repos/)(.*)
>> /pulp/content/var/www/pub/yum/https/repos/pulp/repos/$2 [DPI]
>>   RewriteCond %{HTTPS} off
>>   RewriteRule (.+/pulp/repos/)(.*)
>> /pulp/content/var/www/pub/yum/http/repos/pulp/repos/$2 [DPI]
>>
>> Next, I have nothing under /var/www/streamer/
>>
>> [root at slave repos]# cd /var/www/streamer/
>> [root at slave streamer]# ls -al
>> total 0
>> drwxr-xr-x. 2 apache apache  6 Oct 11 14:33 .
>> drwxr-xr-x. 5 root   root   60 Oct 30 12:07 ..
>>
>>
>> I see the requests for the package coming in from the squid access log on
>> the "slave" pulp server
>>
>> [root at slave streamer]# tail -f /var/log/squid/access.log
>> 1540919365.408      0 127.0.0.1 TAG_NONE/400 4221 GET
>> /var/lib/pulp/content/units/rpm/ea/f2480dc9740f9a787e69e5a7d06f501a67780672f2a35ed22c182e7be387bc/puppet_test_package-1.0.2-1.el6.x86_64.rpm
>> - HIER_NONE/- text/html
>>
>> That content obviously does not exist because it hasn't been downloaded
>> yet..
>>
>> I see under the published directory the link to the non-existent content
>> unit
>>
>> [root at slave p]# pwd
>>
>> /var/lib/pulp/published/yum/https/repos/pulp/repos/rhel-x86_64-server-7-base-thd-tools/Packages/p
>> [root at slave p]# ll
>> total 0
>> lrwxrwxrwx. 1 apache apache 140 Oct 30 13:01
>> puppet_test_package-1.0.2-1.el6.x86_64.rpm ->
>> /var/lib/pulp/content/units/rpm/ea/f2480dc9740f9a787e69e5a7d06f501a67780672f2a35ed22c182e7be387bc/puppet_test_package-1.0.2-1.el6.x86_64.rpm
>>
>> Any ideas?
>>
>>
>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/pulp-list/attachments/20181106/cd0bd116/attachment.htm>