Red Hat Bugzilla – Bug 853025
Monitor scout push failed, after upgrade Sat5.3.0 -> Sat5.5.0
Last modified: 2012-10-02 06:54:41 EDT
Created attachment 608105 [details]
Output by restarting satellite
Description of problem:
I have got this message (attachment) by the last step by upgrade Sat5.3.0 -> Sat5.5.0 embedded DB.
Monitor scout push failed in WebUI.
I tried to re-push, but no change.
In log, I found HTTP error 403.
GET /satconfig/cgi-bin/fetch_netsaintid.cgi?ssk=b503673991b6&publickey=ssh-dss%20AAAAB3NzaC1kc3MAAACBANJzZd6uwHpTScEaNo9gYSjkcrqXrNS4vwcNVvUVX11AqpP8b0Xo4Jm9hiYjz0giSWzfEC474M78T3jEmG10YoQXFZ8GpEBcfnoQeKnocVyFhYMp%2Bre011DDBU%2Frj2gaGBz%2FdAatj6yYbuXHlsgvdFZBhDG4MXwYhv74Ixro0MYnAAAAFQCRAU0maWciF4GEmii5dl%2BsYjs%2FaQAAAIEAygVRNtpWp1dAU8fdiU%2BT3lACP%2F9vSAmHaotdDC56Rh0BBpEcJjiPfWCI0bfKxmir0v5xleSKP6nq%2Fov%2FsRVj4TGFlr%2Fwg3wPYvp8pHEe%2B0AC7w085snr0qKzZVETuMEnAECeaVCpnKlorOhCqyzOtF6kr8%2Fuw6EtiFBXssnqIVQAAACAWLkrLGvfr3Kp0%2BCR0SFgwK%2FBpbAPoI8yGmoy8gAITVaHFYUXWScF7bqS%2Bwn9vrw1bLI4EsE1ABQ66Kz7tg86kfTLPEJ4iLLnuu0KaAlYrtXNTgfAPyGLqoLIdvR0vzIXtnIodP%2F2DmUTrQGk1didnzvWbQG5v1hRogRShUPnruc%3D%20nocpulse%40intel-s3ea2-02%2Erhts%2Eeng%2Ebos%2Eredhat%2Ecom%0A HTTP/1.1" 403 3642 "-" "AgentName/0.1 libwww-perl/5.833
POST /cgi-bin/https_mx.cgi HTTP/1.1" 403 3642 "-" "libwww-perl/5.833
GET /satconfig/cgi-mod-perl/fetch_commands.cgi?cluster_id=&node_id=&role=lead&version=1.0 HTTP/1.1" 403 3642 "-" "libwww-perl/5.833
I tried disable & re-enable "Enable Monitoring Scout" by WebUI, but no change.
In log, I found HTTP error 503.
GET /cgi-bin/translate_key.cgi?scout_shared_key= HTTP/1.1" 503 382 "-" "SatIDXL8r/1.0 libwww-perl/5.833
Whole file /var/log/httpd/access_log is in attachment.
Version-Release number of selected component (if applicable):
Created attachment 608112 [details]
The reason for that 403 on fetch_netsaintid.cgi seems to be in monitoring/eventReceivers/MonitoringAccessHandler.pm -- if the client IP address is not in
my $cf_db = NOCpulse::CF_DB->new;
my $node_ref = $cf_db->getNodes;
The 503 on the translate_key.cgi is IMO caused by the missing key value ... which is likely related to the other calls failing with 403.
I suspect the machine was not just upgraded from 5.3 but migrated to different machine as well.
In that case, spacewalk-hostname-rename probably needs to be run to update the IP address as well ...
I assume no regression thou.
Fixed in Spacewalk master, a29ad894e3ec4db369414b02130bf14bf4f4ebdb.