Subscription-manager never returns?

Hi all,

I am running katello 3.0RC4, which is running Foreman 1.11.1. Have any of
you seen subscription-manager fail to return when trying to register
systems? E.g.

[root@wellminiondev ~]# subscription-manager unregister
System has been unregistered.
[root@wellminiondev ~]# subscription-manager register --org="NIWA"
–activationkey="Apache Production Servers"
The system has been registered with ID: e005c6ff-2494-474b-ac7f-8151548b8d2d

User interrupted process.
[root@wellminiondev ~]# subscription-manager status

··· ----------------------------------------- System Status Details ----------------------------------------- Overall Status: Current

[root@wellminiondev ~]#

As you can see after I ctrl^c the process, the client is registered and its
status is current. BUT if I try complex command than “status” it also hangs

[root@wellminiondev ~]# subscription-manager repos


User interrupted process.

[root@wellminiondev ~]#

and I eventually have to ctrl^c that as well. While for the above I could
ctrl^c the process, during imaging this doesn’t work and the kickstart
process hangs at “Registering the system”. I’ve logged Bug #15019: Katello 3.0RC4 registering systems never returns - Katello - Foreman
which has the output of /var/log/foreman/production.log for a system that
is being kickstarted

Any ideas?


I changed the Katello server VM resources from 8GB of RAM to 12GB and it is
now working - not really sure that's a great answer to the problem tho :slight_smile:

Now, after a "yum update" on the host, it is stuck at "Uploading package

From /var/log/foreman/production.log

2016-05-12 16:31:29 [app] [I] Completed 200 OK in 347ms (Views: 0.6ms |
ActiveRecord: 28.8ms)
2016-05-12 16:32:01 [app] [I] Started PUT
"/rhsm/consumers/d7daa90e-9108-4f93-b7f5-7382bc57e148/packages" for at 2016-05-12 16:32:01 +1200
2016-05-12 16:32:02 [app] [I] Processing by
Katello::Api::Rhsm::CandlepinProxiesController#upload_package_profile as
2016-05-12 16:32:02 [app] [I] Parameters: {"_json"=>[{"vendor"=>"Fedora
Project", "name"=>"zeromq", "epoch"=>0, "version"=>"4.0.5",
"release"=>"4.el7", "arch"=>"x86_64"},

"arch"=>"x86_64"}, {"vendor"=>"Fedora Project", "name"=>"libtommath",
"epoch"=>0, "version"=>"0.42.0", "release"=>"4.el7", "arch"=>"x86_64"},
{"vendor"=>"CentOS", "name"=>"rpm-python", "epoch"=>0, "version"=>"4.11.3",
"release"=>"17.el7", "arch"=>"x86_64"}, {"vendor"=>nil, "name"=>"PyYAML",
"epoch"=>0, "version"=>"3.11", "release"=>"1.el7", "arch"=>"x86_64"},
{"vendor"=>"CentOS", "name"=>"libXau", "epoch"=>0, "version"=>"1.0.8",
"release"=>"2.1.el7", "arch"=>"x86_64"}, {"vendor"=>"CentOS",
"name"=>"python-urllib3", "epoch"=>0, "version"=>"1.10.2",
"release"=>"2.el7_1", "arch"=>"noarch"}, {"vendor"=>"CentOS",
"name"=>"colord-libs", "epoch"=>0, "version"=>"1.2.7", "release"=>"2.el7",
"arch"=>"x86_64"}], "id"=>"d7daa90e-9108-4f93-b7f5-7382bc57e148"}

and then nothing

And solved (I think)
my dev instance was not running the same backup script as my production, I
moved that over today and that includes a vacuum of the postgres databases.
After running the script, everything has come right!
I think specifically its this part that has done it


databases=(foreman candlepin postgres)

Vacuum databases (housekeeping!)

for db in "${databases[@]}"
timeinfo=date '+%x %T'
echo "$timeinfo : $db vacuuming started" >> $logfile
/usr/sbin/runuser - postgres -c "/usr/bin/vacuumdb -z -U postgres
$db" 2>> $logfile
timeinfo=date '+%x %T'
echo "$timeinfo : Vacuum completed successfully on $db" >> $logfile