Created attachment 502790 [details] RHN-Proxy 541 Logs for the Provisioning process. Description of problem: Provisioning of a Host/Guest via RHN-Proxy 541 fails. The provisioning process stops on the first Screen : Choose a Language. I guess the kickstart is not downloaded. I see the following logs on the RHN-Proxy. ### /var/log/squid/access.log ########################################## 1307114065.060 514 127.0.0.1 TCP_MISS/200 331 GET http://127.0.0.1/download//cblr/svc/op/ks/system/M23-Proxy-x86_64-RHEL6-via-M22_060311-09-32-06:1 - DIRECT/127.0.0.1 text/plain 1307114316.145 702 127.0.0.1 TCP_MISS/200 331 GET http://127.0.0.1/download//cblr/svc/op/ks/system/M23-Proxy-x86_64-RHEL6-via-M22_060311-09-32-06:1 - DIRECT/127.0.0.1 text/plain ### /var/log/squid/cache.log ########################################## 2011/06/03 11:14:25| Exception error:corrupted chunk size 2011/06/03 11:18:36| Exception error:corrupted chunk size I tried with both P/E SELinux. Version-Release number of selected component (if applicable): Satellite 541 RHN-Proxy 541 How reproducible: Always Steps to Reproduce: 1. Register Host via RHN-Proxy 541 RHEL6 to Satellite 541 RHEL6 2. Set up Provisioning for the Host via RHN-Proxy Actual results: The provisioning process stacks. Expected results: Success. Additional info: All logs are attached.
It only happen if you use ssl with proxy. If you use proxy.use_ssl=0, then this problem disappear, which is nice for debugging :( From comunication between redirect module and squid I got: HTTP/1.1 200 OK Date: Mon, 06 Jun 2011 19:55:17 GMT Server: Apache/2.2.3 (Red Hat) transfer-encoding: chunked connection: Keep-Alive, close content-type: text/plain; charset=UTF-8 keep-alive: timeout=15, max=400 # Kickstart config file generated by RHN Satellite Config Management # Profile Label : RHEL6-x86_64 # Date Created : 2011-05-30 11:24:44.0 ... The problematic is: transfer-encoding: chunked It should not be there. But right now I do not know who add it there. Can be satellite, apache on proxy or our code in redirector...
This only happen if parent of RHN Proxy is Satellite 541@RHEL6. It does not matter if proxy is 540 or 541. I verified that it does not happen with RHN Proxy with parent Satellite 541@RHEL5.
If proxy get chunked data, httplib will join them so it is not correct to forward transfer-encoding: chunked header. Fixed in spacewalk.git in commit 637ac41756910ef7808d869aa86f520e0206de05 Now we are able to dowload ks file, but stil having problem with downloading rpm files.
The main trouble was caused by sitemesh, that ignored Content-Length header (even if it was manually set) in case Content-Type wasn't set. No bueno. spacewalk.git: a5e30664d5780c078e3a51ccc3c97cbaf275cb7a
Technical note added. If any revisions are required, please edit the "Technical Notes" field accordingly. All revisions will be proofread by the Engineering Content Services team. New Contents: Cause: Incorrect HTTP response headers Consequence: Kickstart over RHN Proxy from a RHN Satellite 5.4.1 on RHEL6 fails Fix: Correct handling of HTTP response headers Result: Kickstart over RHN Proxy from a RHN Satellite 5.4.1 on RHEL6 works now
Verified Satellite-541 RHEL5/i386 & RHN-Proxy RHEL5(i386,x86_64) & RHN-Proxy RHEL6(x86_64) Satellite-541 RHEL6/x86_64 & RHN-Proxy RHEL[56]/x86_64
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. http://rhn.redhat.com/errata/RHBA-2011-0943.html
*** Bug 824140 has been marked as a duplicate of this bug. ***