46 reputation
4
bio website
location
age
visits member for 3 years, 3 months
seen Aug 15 at 0:53

(my about me is currently blank)

click here to edit


Sep
24
awarded  Autobiographer
Apr
26
asked Is there any way to backup the content of business google app presence to local storage?
Apr
2
awarded  Supporter
Mar
14
comment How do I test the effect of ionice (against a device using the cfq scheduler)?
I tried running two instances of my test app (one ionice'd to idle), each in its own window. I started up a stress -d 1 in a third window. Prior to starting stress, I got similar (and marked) progress from each test instance. After starting stress, both slowed appreciably but made essentially identical progress. I tried running stress ionice'd to idle and observed the same: even the idle stress slowed down both test app instances similarly (I would have expected the best-effort instance to have been at least less affected).
Mar
14
comment How do I test the effect of ionice (against a device using the cfq scheduler)?
OK, I'll give it a try-- though at first glance, stress appears to be doing what I'm doing (but in a less deterministic fashion)-- stress -i simply calls sync(2) in quick succession, which I think may or may not induce I/O, depending on other activity in the system -- unlike my attempt, which always makes sure there is something "dirty" to be flushed via fsync(2). Similarly, stress -d appears to be doing a bunch of write(2) operations --which I think can be buffered in the kernel until it calls close(2).
Mar
14
comment How long is the “defined grace period” for the idle scheduling class of the CFQ io scheduler?
The more general question is over at serverfault.com/q/369382/95287
Mar
13
asked How do I test the effect of ionice (against a device using the cfq scheduler)?
Mar
1
comment How long is the “defined grace period” for the idle scheduling class of the CFQ io scheduler?
The more I play around with this, the more I'm convincing myself that the grace period is zero. I'm on FC7 and I've tried writing some I/O intensive processes and assigning them to different scheduling classes . No matter how much I saturate the cfq-scheduled block device, I never see any blocking on the "idle" processes. I think maybe it's time for me to ask a more general question of "how do you test that ionice is actually working".
Feb
24
comment How long is the “defined grace period” for the idle scheduling class of the CFQ io scheduler?
If I'm understanding that RH page correctly, is it not actually slice_idle (and thus 8 ms)?
Feb
24
comment How long is the “defined grace period” for the idle scheduling class of the CFQ io scheduler?
I'm trying to understand what the grace period is so that I can figure out if it makes sense for the load that I'm putting on a box.
Feb
24
comment How long is the “defined grace period” for the idle scheduling class of the CFQ io scheduler?
linux (any distribution with the CFQ io scheduler).
Feb
24
asked How long is the “defined grace period” for the idle scheduling class of the CFQ io scheduler?
Oct
29
awarded  Student
Sep
27
awarded  Tumbleweed
Sep
20
asked preventing tmpwatch from clobbering archives that were “just” unpacked into /tmp