Description of problem:
Kickstarting a RHEL5 machine through a RHN Proxy 4.2.3 is taking an unusually long time in all regions. The installation will 'pause' at "please wait a moment" screen. This is when we see in the proxy logs that the partial contents are not cached and these partial content requests are retrieved from the satellite server itself. I understand that the version of squid running does not support(does not cache) partial content requests.
It is only when all these partial contents are retrieved from the satellite server, will the installation request for the package itself. This is also when the installation screen changes and you will see the progress bars. When the package itself is requested (not partial content) then we see a TCP_HIT in the proxy logs. From here on the installation is back to normal, in other words, fast.
Version-Release number of selected component (if applicable):
RHN Satellite 4.2.x
RHN Proxy 4.2.x
Steps to Reproduce:
1. Install a RHEL4 system and setup a RHN Proxy 4.2.x
1.Kickstart a RHEL5 system through the RHN Proxy
2.Monitor the time it takes to fetch the packages
3.Re-kickstart the same system again and see the time it takes to fetch packages
NO caching happens
The second install should take less time than the first one.
This does not effect RHEL3/4 kickstarts. This doesn't happen on newer RHNS (530)
Cherry-picked to satellite.git as commit 520c280af9227cf189f23a3d466400774ae857d6
Changing summary as the fix is for 5.3
An advisory has been issued which should help the problem
described in this bug report. This report is therefore being
closed with a resolution of ERRATA. For more information
on therefore solution and/or where to find the updated files,
please follow the link below. You may reopen this bug report
if the solution does not work for you.