https://bugzilla.novell.com/show_bug.cgi?id=740631 https://bugzilla.novell.com/show_bug.cgi?id=740631#c2 --- Comment #2 from Scott Couston 2012-01-10 21:56:26 UTC --- Created an attachment (id=470568) --> (http://bugzilla.novell.com/attachment.cgi?id=470568) Image 2 - Note that further resorces arre being devoted to cache and the available RAM cannot support/sustain opebing of ANY other Application The other ugly...As the Physical RAM is almost exhausted, why has the kernel NOT chosen to use Virtual RAM - In both images there is almost NO RAM left but ALL available Virtual RAM has not even been touched or used at this far too late, critical level -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=740631 https://bugzilla.novell.com/show_bug.cgi?id=740631#c4 --- Comment #4 from Scott Couston 2012-01-10 22:06:05 UTC --- Created an attachment (id=470570) --> (http://bugzilla.novell.com/attachment.cgi?id=470570) Different PC - Same Problem On this PC the only applications that I have open are 1. A Download of opensuse 12.1 - That's chewing up ALL the cache Buffers, Firefox and Thunderbird - Thats ALL...Currently the kde Window manager is so so so slow and I can out run the presentation of text using 1 finger..Its that criminally slow!!! -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=740631 https://bugzilla.novell.com/show_bug.cgi?id=740631#c6 --- Comment #6 from Scott Couston 2012-01-10 22:23:53 UTC --- Perhaps you can help me understand...What time or otherwise parameter determines a cache buffer to be dirty AND what time or other parameter flushes dirty cache buffers and return the RAM allocation to the Pool...Currently I have 49 Meg of physical RAM available to run applications and the Windowing Manager, out of 4GIG -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=740631 https://bugzilla.novell.com/show_bug.cgi?id=740631#c8 --- Comment #8 from Scott Couston 2012-01-11 20:19:52 UTC ---

Why do you think this is a RAM issue at all? The kernel never uses up cache memory at the expense of an application, see the page I linked to above for details about that, and how it all works.

In the pics I have sent you the available RAM gets down to almost 128Meg out of 4 GIG. Its clearly visible. When the available RAM gets lower than 128Meg the O/S falls over. I can leave a PC that has used 3.6GIG of RAM devoted to Cache, come back in an hour and ache has not been flushed. This is very easily duplicated by using ANY multimedia application and a DVD full of many many large files...With less that 128Meg of RAM available in both my pics the O/S predictably falls over..Backtrace to be uploaded as soon as possible, however the backtrace is NOT going to tell your why the PC fell Over as the root cause is not the application or Windowing Manager falling over because of a code problem...This is why this problem has so many bugs open surrounding this issue but not providing accurate backtrace info to solve the basic problem. There are also bugs created against gnome and other Multimedia applications that fall over as the number of files and the size of the files to cache from DVD is so huge! Can you tell me which audit logs I can upload immediately that contains the history of the Windowing Manager or O/S falling over..thanks -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=740631 https://bugzilla.novell.com/show_bug.cgi?id=740631#c10 --- Comment #10 from Scott Couston 2012-01-31 19:10:04 UTC --- Created an attachment (id=473592) --> (http://bugzilla.novell.com/attachment.cgi?id=473592) 1 of 5 in correct time line order I am happy to leave this bug closed but I'm going to add some thoughts that might just help later on. My issue is the DYNAMIC use and allocation of cache buffers. After a DVD is inserted with many files the cache buffers are allocated to use almost 90% of base RAM. If the same PC is left alone the cache buffers remain and the dynamic allocation of the ceases to function. After 90% of the total system RAM is allocated to cache buffers and considering they are not dynamically flushed even after 3 days has gone by there IS a problem. In the following example the PC has 4GB of RAM and I have an application running that has the kernel dynamically allocates 90% of base RAM to cache buffers. When I go to use Yast NFS Client the windowing manager freezes because the kernel has not dynamically removed enough RAM from cache buffers for the application to complete. In fact in this test condition with 90% of RAM being used for cache buffers, many applications wont even launch due the lack of base RAM. The dynamic removal or reallocation of RAM used as cache buffers is not functioning. Some aspect of the kernel has changed where the exists NO dynamic allocation of cache buffers, NO dynamically added flag setting some cache buffers as dirty and no dynamically controlled flushing of the dirty cache buffers and dynamically returned to base RAM for further applications to launch and/or not freeze and not slow due to the 10% being base RAM so the PC doesn't crawl and/or not complete its task and not even opening some large aps like ooo. For the record I don’t open kernel bugs unless I am absolutely sure of the problem and the problem is 100% reproducible on my quad core X64 AMD PC. I am not in the habit of creating work for the developers that entails log audit logs to be analysed or in this case the kernel being changed and recompiled. Dynamic allocation of RAM is a complex algorithm, I know this but I am sure something in the current Kernel is stopping the process of dynamic use of RAM and debugging this part is hugely time consuming and not to be taken lightly - I still believe the bug still exists, so for future posterity I'll leave it closed -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=740631 https://bugzilla.novell.com/show_bug.cgi?id=740631#c12 --- Comment #12 from Scott Couston 2012-01-31 19:11:30 UTC --- Created an attachment (id=473594) --> (http://bugzilla.novell.com/attachment.cgi?id=473594) 3 of 5 in correct time line order -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=740631 https://bugzilla.novell.com/show_bug.cgi?id=740631#c14 --- Comment #14 from Scott Couston 2012-01-31 19:12:41 UTC --- Created an attachment (id=473596) --> (http://bugzilla.novell.com/attachment.cgi?id=473596) 5 of 5 in correct time line order -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=740631 https://bugzilla.novell.com/show_bug.cgi?id=740631#c16 Scott Couston changed: What |Removed |Added ---------------------------------------------------------------------------- Summary|Total System Crash - |Error in the DYNAMIC |Irrespective of Software |allocation of Physical RAM |Version - Applications Fail |where Apps wont launch and |to Launch - Windowing |SWAP is NOT utilised |Manager Crashes - Signal | |11 Fault | --- Comment #16 from Scott Couston 2012-02-05 09:52:38 UTC --- Greg, I thank you for the URL reference to how RAM is used in C #3, but this itself begs the question of why does the web page need to exist at all if everything about RAM was functioning perfectly and there are sufficient numbers of users having problems in the fist instance for the web page to exist at all! I have apps that wont launch when avail ram is at 84MG out of a total of 4GIG. I have stored images showing that even when available RAM sinks as low as 84Meg, swap or virtual RAM is not utilised what-so-ever. Then I have the problem that after three days of not touching the PC, the cache RAM allocation remains constant 3 days after the removable media, that was cached, was ejected. When are cache buffers deemed to be dirty and then flushed - certainly in my test case the files being cached were on a removable drive and the DVD was ejected, but the cache buffers that no longer existed were not marked dirty after three days. Your web page happily says that if an application asks for RAM that was previously dedicated to cache, that it would be dynamically given RAM - The problem is the DYNAMIC allocation of RAM between cache and apps just is not happening for so many users to complain or raise questions! It also appears that DYNAMIC use of virtual or swap RAM does not occur. It also appears that there is NO parameter that flags cache ram as being dirty and following that there in no DYNAMIC flush of Dirt Cache buffers! I am happy with your expertise that is greater than mine; but it begs the question that if all is well, why was a web page created to explain to so many users that their concerns were not valid. There was a time I can recall that you could not ever see serious amounts of RAM being Dynamic allocated to cache. It was horrible to see that despite huge available amounts of RAM very little was devoted to cache...Now we seem to have gone overboard in this direction with apparent dire consequences. If nothing I am happy to leave as invalid as I was serious when I said your knowledge exceeds mine, however my I suggest a unique hit counter on your web explanation page as growing hit numbers cannot rationally be ignored..thanks mate:-) -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=740631 https://bugzilla.novell.com/show_bug.cgi?id=740631#c Scott Couston changed: What |Removed |Added ---------------------------------------------------------------------------- AssignedTo|kde-maintainers@suse.de |kernel-maintainers@forge.pr | |ovo.novell.com OS/Version|openSUSE 11.4 |openSUSE 12.2 Severity|Normal |Major -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=740631 https://bugzilla.novell.com/show_bug.cgi?id=740631#c18 Scott Couston changed: What |Removed |Added ---------------------------------------------------------------------------- Target Milestone|--- |RC 1 OS/Version|openSUSE 12.1 |openSUSE 12.2 --- Comment #18 from Scott Couston 2012-02-10 03:26:16 UTC --- Jiri, Please chance the viability of this bug to Novell/Suse.de only - You know I would not ask unless this were importing. Scott -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.