Hey,
I tried to use CFQ and ionice to help some I/O intensive tasks on a
MySQL slave, and it turned out to makes thing much worse to the point I
can figure out what it could be beside a bug. Tested on 2.6.20.1, but I
could eventually upgrade if there's CFQ bugfixes I'm missing. The
filesystem is ReiserFS.
When "idle", the slave do mostly random writes: about 15 rtps and 200
wtps. For testing I ended up running dd to read files from disk to
/dev/null while avoiding the system cache (had 42GB of data to read
from, and <1GB of ram cache/buffers). Here's some sample results (I
tried them multiple times with similar results every time). I monitored
the iops with "sar -b 1 0"
Under deadline-iosched:
262144000 bytes (262 MB) copied, 9.68511 seconds, 27.1 MB/s
rtps raised over 200, wtps gets down to ~100 and then back up after the
read operation (as MySQL catch up)
Under cfq-ipoched without ionice:
262144000 bytes (262 MB) copied, 4.78834 seconds, 54.7 MB/s
rtps raise over 400, wtps gets down near 0 then back up after the read.
I can expect it as cfq does not manage the write queue, so it gives most
bandwidth to dd
Under cfq-ipoched with ionice -c3 (mysqld was "best-effort: prio 4"):
13369344 bytes (13 MB) copied, 39.7619 seconds, 336 kB/s (After pressing
CTRL-C to avoid tripping off alerting on MySQL replication)
Mysql was lagging behind at pretty much the same rate with and without
"ionice -c3", but with the latter the copy operation was also incredibly
slower. During the read operation rtps was around 3 and wtps was between
10 and 20, far beyond what that raid array is able to do with pure
random operations.
Any ides what's going on with this scheduler?
FWIW the dd command was:
dd if=<file> of=/dev/null bs=128k skip=NNNN count=2000
Where NNNN is a multiple of 2000 incremented by 1 on every run.
Thanks,
Thomas
> On Thu, 09 Oct 2008 18:34:06 -0400 Thomas Guyot-Sionnest <[email protected]> wrote:
> Hey,
>
> I tried to use CFQ and ionice to help some I/O intensive tasks on a
> MySQL slave, and it turned out to makes thing much worse to the point I
> can figure out what it could be beside a bug. Tested on 2.6.20.1, but I
> could eventually upgrade if there's CFQ bugfixes I'm missing. The
> filesystem is ReiserFS.
>
> When "idle", the slave do mostly random writes: about 15 rtps and 200
> wtps. For testing I ended up running dd to read files from disk to
> /dev/null while avoiding the system cache (had 42GB of data to read
> from, and <1GB of ram cache/buffers). Here's some sample results (I
> tried them multiple times with similar results every time). I monitored
> the iops with "sar -b 1 0"
>
> Under deadline-iosched:
>
> 262144000 bytes (262 MB) copied, 9.68511 seconds, 27.1 MB/s
> rtps raised over 200, wtps gets down to ~100 and then back up after the
> read operation (as MySQL catch up)
>
> Under cfq-ipoched without ionice:
>
> 262144000 bytes (262 MB) copied, 4.78834 seconds, 54.7 MB/s
> rtps raise over 400, wtps gets down near 0 then back up after the read.
> I can expect it as cfq does not manage the write queue, so it gives most
> bandwidth to dd
>
> Under cfq-ipoched with ionice -c3 (mysqld was "best-effort: prio 4"):
> 13369344 bytes (13 MB) copied, 39.7619 seconds, 336 kB/s (After pressing
> CTRL-C to avoid tripping off alerting on MySQL replication)
>
> Mysql was lagging behind at pretty much the same rate with and without
> "ionice -c3", but with the latter the copy operation was also incredibly
> slower. During the read operation rtps was around 3 and wtps was between
> 10 and 20, far beyond what that raid array is able to do with pure
> random operations.
>
> Any ides what's going on with this scheduler?
>
> FWIW the dd command was:
>
> dd if=<file> of=/dev/null bs=128k skip=NNNN count=2000
> Where NNNN is a multiple of 2000 incremented by 1 on every run.
>
Yes, 2.6.20 is dreadfully old. If you can retest a recent kernel it would really help, thanks.
On 14/10/08 03:18 PM, Andrew Morton wrote:
>> On Thu, 09 Oct 2008 18:34:06 -0400 Thomas Guyot-Sionnest <[email protected]> wrote:
>> Hey,
>>
>> I tried to use CFQ and ionice to help some I/O intensive tasks on a
>> MySQL slave, and it turned out to makes thing much worse to the point I
>> can figure out what it could be beside a bug. Tested on 2.6.20.1, but I
>> could eventually upgrade if there's CFQ bugfixes I'm missing. The
>> filesystem is ReiserFS.
>>
>> When "idle", the slave do mostly random writes: about 15 rtps and 200
>> wtps. For testing I ended up running dd to read files from disk to
>> /dev/null while avoiding the system cache (had 42GB of data to read
>> from, and <1GB of ram cache/buffers). Here's some sample results (I
>> tried them multiple times with similar results every time). I monitored
>> the iops with "sar -b 1 0"
>>
>> Under deadline-iosched:
>>
>> 262144000 bytes (262 MB) copied, 9.68511 seconds, 27.1 MB/s
>> rtps raised over 200, wtps gets down to ~100 and then back up after the
>> read operation (as MySQL catch up)
>>
>> Under cfq-ipoched without ionice:
>>
>> 262144000 bytes (262 MB) copied, 4.78834 seconds, 54.7 MB/s
>> rtps raise over 400, wtps gets down near 0 then back up after the read.
>> I can expect it as cfq does not manage the write queue, so it gives most
>> bandwidth to dd
>>
>> Under cfq-ipoched with ionice -c3 (mysqld was "best-effort: prio 4"):
>> 13369344 bytes (13 MB) copied, 39.7619 seconds, 336 kB/s (After pressing
>> CTRL-C to avoid tripping off alerting on MySQL replication)
>>
>> Mysql was lagging behind at pretty much the same rate with and without
>> "ionice -c3", but with the latter the copy operation was also incredibly
>> slower. During the read operation rtps was around 3 and wtps was between
>> 10 and 20, far beyond what that raid array is able to do with pure
>> random operations.
>>
>> Any ides what's going on with this scheduler?
>>
>> FWIW the dd command was:
>>
>> dd if=<file> of=/dev/null bs=128k skip=NNNN count=2000
>> Where NNNN is a multiple of 2000 incremented by 1 on every run.
>>
>
> Yes, 2.6.20 is dreadfully old. If you can retest a recent kernel it would really help, thanks.
Did it, although I had to patch it as the AACRAID driver is broken since
2.6.25! See:
http://bugzilla.kernel.org/show_bug.cgi?id=9133
https://bugzilla.redhat.com/show_bug.cgi?id=453472
https://bugzilla.redhat.com/show_bug.cgi?id=457552
http://marc.info/?l=linux-kernel&m=122166454808377&w=2
With 2.6.27.2 the Idle class performs much better now, and using it, my
dd command has less impact on the MySQL replication than using the
default class.
Interestingly though, the CFQ Idle class it still clearly having more
impact on replication than using the deadline scheduler. Anyone
interested in more details (I compiled the kernel with a bunch of IO
scheduler statistics in case it would be needed).
Thomas
Thomas Guyot-Sionnest wrote:
> On 14/10/08 03:18 PM, Andrew Morton wrote:
>>> When "idle", the slave do mostly random writes: about 15 rtps and 200
>>> wtps. For testing I ended up running dd to read files from disk to
>>> /dev/null while avoiding the system cache (had 42GB of data to read
Just thought I would point out that dd is, in fact, dirtying the buffer
cache since you don't pass flag=direct.
On 28/10/08 02:07 PM, Phillip Susi wrote:
> Thomas Guyot-Sionnest wrote:
>> On 14/10/08 03:18 PM, Andrew Morton wrote:
>>>> When "idle", the slave do mostly random writes: about 15 rtps and 200
>>>> wtps. For testing I ended up running dd to read files from disk to
>>>> /dev/null while avoiding the system cache (had 42GB of data to read
>
> Just thought I would point out that dd is, in fact, dirtying the buffer
> cache since you don't pass flag=direct.
>
Yes, I know that, though at each run I read a different chunk so it
shouldn't affect my benchmark.
The point to note is that even though the CFQ Idle class seems to work
now (it does reduce the IO load instead of making overall IO a lot
slower), it's sill having more impact on my system than using the
deadline scheduler which has no class at all.
--
Thomas