Age | Commit message (Collapse) | Author |
|
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
|
|
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
|
|
This reverts commit f03f8e473b7c8a5c4ed2de0b99717cf75521bf69.
|
|
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
|
|
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
|
|
There were issues in having the same underlying device being referenced
multiple times (via different paths) when reporting storage I/O
statistics. As an example: having two (or more) LVM2/DM volumes crafted
out of the same MD array.
This patch simply skips over any devices previously seen.
Signed-off-by: Alan D. Brunelle <alan.brunelle@hp.com>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
|
|
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
|
|
If a job file contains more than one job that works on separate
disks, then the job that finishes first will get more idle time
accounted to it.
Fix this by only doing disk util stats when we have files open
on that device.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
|
|
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
|
|
But do so by adding new fields in the stats line.
We dont want the existing output format to get clobbered.
The new fields in master device line which contain
aggregated stats start with "aggr".
Signed-off-by: Shehjar Tikoo <shehjart@gelato.unsw.edu.au>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
|
|
When running tests on a device created using mdadm, the relevant
disk statistics must also come from the individual
disks that constitute the RAID, because the software RAID device
is just a virtual device.
This patch goes through the 'slaves' directory in the software
RAID device's directory in /sys/block, gets a list of all the
slave devices and creates a master-slave relationship between
the virtual device and the actual disks.
The individual slave disks' stats are displayed along with the
master's disk stats at the end of the test.
Signed-off-by: Shehjar Tikoo <shehjart@gelato.unsw.edu.au>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
|
|
/sys/block/<dev> is now a sym link to the real path, which
broke our disk lookup. Fix it.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
|
|
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
|
|
It's still the same timer and shared, but move it out of
diskutil so we get proper status reporting on non-Linux.
We also need to rearm the signal handlers on each signal.
Move this to sigaction() next.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
|
|
Caused fio to stall waiting for disk updates.
Also add --debug=diskutil debug option, to trace what diskutil is doing.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
|
|
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
|
|
Add an option to disable the disk stats. And only dump the disk
stat header, if we actually have stats to show.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
|
|
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
|
|
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
|
|
It's only used if FIO_HAVE_DISK_UTIL is set.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
|