Hide Forgot
Created attachment 478973 [details] top output from client machine which was running rhn_check Description of problem: When I schedule say 170 errata to install on a client machine, I notice that the first 80 updates install quickly however after that it takes a least a few minutes for each update to install. I monitored the memory usage for rhn_check and it seems to keep growing and eventually my test client (which is a VMware guest with 512MB of RAM) starts swapping which brings the system to its knees. Version-Release number of selected component (if applicable): rhn_check (Red Hat Network Client Tools) 0.4.19-17.el5 Copyright (C) 1999-2006 Red Hat, Inc. Licensed under the terms of the GPL. How reproducible: Every time I schedule a large job such as the one previously described I can reproduce it. Steps to Reproduce: 1. Schedule at least 100 or so errata to be installed on a client 2. Wait for job to be picked up by client and watch rhn_check's memory footprint grow and grow Actual results: Slow installation of all errata (in my example I was going from RHEL 5.3 to the most current release including the latest errata) Expected results: Limited memory usage and quick installation of all errata (for example when I do a yum update it never takes longer than 10-15 minutes to install this number of updates) Additional info: See attached screenshot of top from the client machine that shows the large memory footprint of rhn_check (It had been running for at least an hour and 15 minutes at this point - 150/173 of the updates were installed)
Can you please review Bug 566746 - I think that this is a duplicate of that issue, which only effects older RHEL 5 systems. Regards, Cliff
Cliff, Thank you for the reply. I have reviewd Bug #566746 however I think this maybe entirely different as we do not run selinux here (I have also confirmed that the test machine had it disabled which it did). Regards, Dan