From Bugzilla Helper: User-Agent: Mozilla/4.77 [en] (X11; U; SunOS 5.6 sun4u) Description of problem: I mounted the RH7.2 iso images via loopback with this command: mount $1 -r -t iso9660 -o loop $2 where $1 is the iso image and $2 is the mount point. All is fine, until I start copying files. I like to copy all the RPMS into a separate directory for easy access. As it copies, the amount of used memory grows and grows until about 5M, and then the system slows down to a crawl. kswapd starts running at about 30% cpu time, and the system load rises into the 9s. Swap memory does not get used during this time either. Eventually, the copy process finishes and system performance and use of resources returns to normal. Note: I did not have this problem with RH7.1 Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1. Mount an ISO image via loopback 2. Start copying data from it 3. When available memory approaches 0, system will slow to a crawl Actual Results: system performance became incredibly slow until the copy process eventually completed Expected Results: Memory should've been freed as the copy proceeded or used swap space as necessary. Additional info: I have 1.5Gb RAM. I successfully reproduced this bug with 500Mb of swap space, and with 2Gb of swap. Turning swap on or off, or changing swap device's priority doesn't seem to affect performance, kswapd still sucks major system resources. I'm running on an Abit KT7A-Raid with a 1.33Ghz Athlon
Is this with the 2.4.7 or the 2.4.9 kernel ? The 2.4.9 kernel should behave better in this case
This was the 2.4.7 kernel. After upgrading to 2.4.9 all is well.
"After upgrading to 2.4.9 all is well." -> I'm closing this as "fixed in errata" as, well, it is. If this comes back or you don't agree with closing as fixed, please reopen this bug.