On 2015-04-22 11:45 PM, Ulrich Windl wrote:
Geoff Swan gswan3@bigpond.net.au schrieb am 22.04.2015 um 14:06 in Nachricht
55378EE0.40103@bigpond.net.au:
On 2015-04-22 4:49 PM, Ulrich Windl wrote:
[...]
Are there any clues about key factors affecting this? Linux, in this case, has vm.swappiness set to 10, vm.dirty_ratio at 12 and vm.dirty_background at 3. However I've noticed that when dirty pages are flushed to disc, the system stalls. And that operation appears to take a relatively long time. Disc write speed should be close to 130MB/s (file copy, dd test etc) however it appears to be much slower than this with the page flush.
Did you try NOT tuning those? A swapped in-memory database is not the thing
you usually want. Swappiness for an out-of-the-box kernel was 60, which sounds way too high. So I reduced it to 10.
Did we see your "free" stats? On a server running SLES11 SP3 (kernel 3.0.101) with similar memory (no slapd, but a huge Oracle and a NFS Server) I have: # free total used free shared buffers cached Mem: 132156276 131202644 953632 40317620 5219804 95059508 -/+ buffers/cache: 30923332 101232944 Swap: 20956756 165852 20790904
I believe what the system pages out is really "dead meat". According to sar there is little actual paging:
00:00:01 pgpgin/s pgpgout/s fault/s majflt/s pgfree/s pgscank/s pgscand/s pgsteal/s %vmeff 11:00:01 12671.34 1979.03 4469.13 0.00 3412.08 0.00 0.00 0.00 0.00 11:10:01 11598.11 1693.74 4949.54 0.01 3466.14 0.00 0.00 0.00 0.00 11:20:01 12141.05 4759.01 4219.09 0.02 3753.33 0.00 0.00 0.00 0.00 11:30:01 14619.11 1562.84 4185.04 0.00 3134.25 0.00 0.00 0.00 0.00
(Most "fage faults" are due to processes staring and ending with demand paging, I guess)
Regards, Ulrich
Free stats look fine to me. No swap is being used, or has been used yet on this system.
total used free shared buffers cached Mem: 132005400 21928192 110077208 10612 363064 19230072 -/+ buffers/cache: 2335056 129670344 Swap: 8388604 0 8388604
The effect I am seeing is that despite a reasonably fast disc system, the kernel writing of dirty pages is painfully slow.