1 |
On Mon, Dec 6, 2010 at 7:58 AM, Nikos Chantziaras <realnc@×××××.de> wrote: |
2 |
> On 12/06/2010 04:33 PM, Stroller wrote: |
3 |
>> |
4 |
>> On 27/11/2010, at 10:22pm, Nikos Chantziaras wrote: |
5 |
>>>> |
6 |
>>>> ... |
7 |
>>>> Does using the ck-sources kernel reduce the necessity to use ionice? |
8 |
>>> |
9 |
>>> To some degree, yes. |
10 |
>> |
11 |
>> Tried this late last week. You chose your words perfectly - there is only |
12 |
>> some degree of improvement. |
13 |
>> |
14 |
>> The machine seems much snappier if I return to an interactive session |
15 |
>> after leaving it idle for some time, but during DVD rips it is still |
16 |
>> unresponsive for tens of seconds at a time. It seems like maybe it responds |
17 |
>> quicker than mainline sources, but it's still so slow that it's both |
18 |
>> unusable and hard to be sure whether that's the case. |
19 |
>> |
20 |
>> Will try cgroups this week, perhaps. |
21 |
> |
22 |
> cgroups will not help with that either. I bet the problem is that the |
23 |
> kernel doesn't care about what kind of data it is caching. I came across |
24 |
> this on LKML, but no one was interested in a solution. |
25 |
> |
26 |
> The problem is that the kernel caches all data. Ripping a DVD will consume |
27 |
> 4GB cache, and all other data will be thrown out. Throwing that all out |
28 |
> again after the rip has finished is an awful lot of I/O load. The smart |
29 |
> thing to do would be for the kernel to not cache such stuff (DVD rips, big |
30 |
> video files, etc.) But it does. |
31 |
|
32 |
I forgot to ask Stroller if he's looked at top while these delays |
33 |
happen. If the issue is writing to disk from cached memory then what I |
34 |
saw when that was occurring (and during the times my system was |
35 |
unresponsive) was very low CPU usage while having very high wait |
36 |
percentages. |
37 |
|
38 |
In my 6 core/12 thread top display I'd see 2 or 3 CPUs involved in the |
39 |
wait (writing data out but delayed by responses from the disk driver I |
40 |
suppose) while the other CPUs were essentially idle, presumably |
41 |
because they couldn't get to disk for whatever they needed. |
42 |
|
43 |
- Mark |