On Wed, Apr 1, 2009 at 10:19 PM, Ayer, Paul W <pwa...@statestreet.com> wrote:
> For swappiness it seems that it would be set by each system and what they > are doing from what I am reading.. Right, to be determined for each system separately, and reviewed when the application or configuration changes. I was just trying to explain why it would not be helpful to yell now "we think 17 is a good number for swappiness" > For the cache I have found that you MUST enter the command sync first then it > all works fine and a free display shows lower after .. without entering the > sync command first they system just hangs up ... very true ... That would be ugly if it would simply drop a page even when dirty and thus hang the system (and hard to believe on any day but today). Obviously between you typing synch and the next command, new dirty pages could be created... You care for some Russian Roulette maybe? Some reading shows there are lockup scenarios when using drop_caches (and the Bugzilla is against RHEL). As for the need to "sync" first: I read the "should sync first" as that it is more effective when you first tell Linux to write out any dirty pages or you would be left later with the clean pages that are still in cache. But as said, when drop_caches does not reduce your memory requirement on z/VM, then I would not know why you would want to do it (apart as diagnostics to understand the baseline requirement for your page cache so that you can compute the swappiness). Rob -- Rob van der Heij Velocity Software http://www.velocitysoftware.com/ ---------------------------------------------------------------------- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390