diff mbox

[KVM-AUTOTEST,4/4] Adding control files dir to kvm test

Message ID 1244823190-9299-4-git-send-email-lmr@redhat.com (mailing list archive)
State New, archived
Headers show

Commit Message

Lucas Meneghel Rodrigues June 12, 2009, 4:13 p.m. UTC 
Adding an autotest_control dir with control files that will be
used on the 'autotest' kvm test, with the original control files
used on the old kvm_runtest_2 directory.

Signed-off-by: Lucas Meneghel Rodrigues <lmr@redhat.com>
---
 client/tests/kvm/autotest_control/bonnie.control   |   21 ++++++++++++++++++++
 client/tests/kvm/autotest_control/dbench.control   |   20 +++++++++++++++++++
 .../tests/kvm/autotest_control/sleeptest.control   |   15 ++++++++++++++
 3 files changed, 56 insertions(+), 0 deletions(-)
 create mode 100644 client/tests/kvm/autotest_control/bonnie.control
 create mode 100644 client/tests/kvm/autotest_control/dbench.control
 create mode 100644 client/tests/kvm/autotest_control/sleeptest.control

Comments

Martin Bligh June 12, 2009, 4:18 p.m. UTC | #1 
That seems very strange ... why do you have to duplicate the existing
control files?

On Fri, Jun 12, 2009 at 9:13 AM, Lucas Meneghel Rodrigues
<lmr@redhat.com> wrote:
> Adding an autotest_control dir with control files that will be
> used on the 'autotest' kvm test, with the original control files
> used on the old kvm_runtest_2 directory.
>
> Signed-off-by: Lucas Meneghel Rodrigues <lmr@redhat.com>
> ---
>  client/tests/kvm/autotest_control/bonnie.control   |   21 ++++++++++++++++++++
>  client/tests/kvm/autotest_control/dbench.control   |   20 +++++++++++++++++++
>  .../tests/kvm/autotest_control/sleeptest.control   |   15 ++++++++++++++
>  3 files changed, 56 insertions(+), 0 deletions(-)
>  create mode 100644 client/tests/kvm/autotest_control/bonnie.control
>  create mode 100644 client/tests/kvm/autotest_control/dbench.control
>  create mode 100644 client/tests/kvm/autotest_control/sleeptest.control
>
> diff --git a/client/tests/kvm/autotest_control/bonnie.control b/client/tests/kvm/autotest_control/bonnie.control
> new file mode 100644
> index 0000000..2717a80
> --- /dev/null
> +++ b/client/tests/kvm/autotest_control/bonnie.control
> @@ -0,0 +1,21 @@
> +AUTHOR = "Martin Bligh <mbligh@google.com>"
> +NAME = "bonnie"
> +TIME = "MEDIUM"
> +TEST_CLASS = "Kernel"
> +TEST_CATEGORY = "Functional"
> +TEST_TYPE = "client"
> +DOC = """\
> +Bonnie is a benchmark which measures the performance of Unix file system
> +operations. Bonnie is concerned with identifying bottlenecks; the name is a
> +tribute to Bonnie Raitt, who knows how to use one.
> +
> +For more info, see http://www.textuality.com/bonnie/
> +
> +This benchmark configuration run generates sustained write traffic
> +of 35-50MB/s of .1MB writes to just one disk.  It appears to have a
> +sequential and a random workload. It gives profile measurements for:
> +throughput, %CPU rand seeks per second. Not sure if the the CPU numbers
> +are trustworthy.
> +"""
> +
> +job.run_test('bonnie')
> diff --git a/client/tests/kvm/autotest_control/dbench.control b/client/tests/kvm/autotest_control/dbench.control
> new file mode 100644
> index 0000000..7fb8a37
> --- /dev/null
> +++ b/client/tests/kvm/autotest_control/dbench.control
> @@ -0,0 +1,20 @@
> +TIME="SHORT"
> +AUTHOR = "Martin Bligh <mbligh@google.com>"
> +DOC = """
> +dbench is one of our standard kernel stress tests.  It produces filesystem
> +load like netbench originally did, but involves no network system calls.
> +Its results include throughput rates, which can be used for performance
> +analysis.
> +
> +More information on dbench can be found here:
> +http://samba.org/ftp/tridge/dbench/README
> +
> +Currently it needs to be updated in its configuration. It is a great test for
> +the higher level I/O systems but barely touches the disk right now.
> +"""
> +NAME = 'dbench'
> +TEST_CLASS = 'kernel'
> +TEST_CATEGORY = 'Functional'
> +TEST_TYPE = 'client'
> +
> +job.run_test('dbench', seconds=60)
> diff --git a/client/tests/kvm/autotest_control/sleeptest.control b/client/tests/kvm/autotest_control/sleeptest.control
> new file mode 100644
> index 0000000..725ae81
> --- /dev/null
> +++ b/client/tests/kvm/autotest_control/sleeptest.control
> @@ -0,0 +1,15 @@
> +AUTHOR = "Autotest Team"
> +NAME = "Sleeptest"
> +TIME = "SHORT"
> +TEST_CATEGORY = "Functional"
> +TEST_CLASS = "General"
> +TEST_TYPE = "client"
> +
> +DOC = """
> +This test simply sleeps for 1 second by default.  It's a good way to test
> +profilers and double check that autotest is working.
> +The seconds argument can also be modified to make the machine sleep for as
> +long as needed.
> +"""
> +
> +job.run_test('sleeptest', seconds = 1)
> --
> 1.6.2.2
>
> _______________________________________________
> Autotest mailing list
> Autotest@test.kernel.org
> http://test.kernel.org/cgi-bin/mailman/listinfo/autotest
>
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Lucas Meneghel Rodrigues June 12, 2009, 4:35 p.m. UTC | #2 
On Fri, 2009-06-12 at 09:18 -0700, Martin Bligh wrote:
> That seems very strange ... why do you have to duplicate the existing
> control files?

In particular, in the old kvm autotest tree we used to have a control
file, control.60 inside the dbench directory, that specifies a shorter
(60s) run for dbench (cause the default 600s is too lengthy for the
large amount of tests we run). 

Instead of sending a patch creating it again inside the dbench test, I
figured a separate dir with the control files we are wanting to run on
the guests a better idea, cause we might want to run fancier control
files on guests. Having the user specify the contents of the control
file on our kvm_tests.cfg is also an option, but then we would need to
handle better multi-line contents on our kvm_config configuration
parser.

All in all, in order to get the old functionality OK I thought having
this extra dir would be the best option.

> On Fri, Jun 12, 2009 at 9:13 AM, Lucas Meneghel Rodrigues
> <lmr@redhat.com> wrote:
> > Adding an autotest_control dir with control files that will be
> > used on the 'autotest' kvm test, with the original control files
> > used on the old kvm_runtest_2 directory.
> >
> > Signed-off-by: Lucas Meneghel Rodrigues <lmr@redhat.com>
> > ---
> >  client/tests/kvm/autotest_control/bonnie.control   |   21 ++++++++++++++++++++
> >  client/tests/kvm/autotest_control/dbench.control   |   20 +++++++++++++++++++
> >  .../tests/kvm/autotest_control/sleeptest.control   |   15 ++++++++++++++
> >  3 files changed, 56 insertions(+), 0 deletions(-)
> >  create mode 100644 client/tests/kvm/autotest_control/bonnie.control
> >  create mode 100644 client/tests/kvm/autotest_control/dbench.control
> >  create mode 100644 client/tests/kvm/autotest_control/sleeptest.control
> >
> > diff --git a/client/tests/kvm/autotest_control/bonnie.control b/client/tests/kvm/autotest_control/bonnie.control
> > new file mode 100644
> > index 0000000..2717a80
> > --- /dev/null
> > +++ b/client/tests/kvm/autotest_control/bonnie.control
> > @@ -0,0 +1,21 @@
> > +AUTHOR = "Martin Bligh <mbligh@google.com>"
> > +NAME = "bonnie"
> > +TIME = "MEDIUM"
> > +TEST_CLASS = "Kernel"
> > +TEST_CATEGORY = "Functional"
> > +TEST_TYPE = "client"
> > +DOC = """\
> > +Bonnie is a benchmark which measures the performance of Unix file system
> > +operations. Bonnie is concerned with identifying bottlenecks; the name is a
> > +tribute to Bonnie Raitt, who knows how to use one.
> > +
> > +For more info, see http://www.textuality.com/bonnie/
> > +
> > +This benchmark configuration run generates sustained write traffic
> > +of 35-50MB/s of .1MB writes to just one disk.  It appears to have a
> > +sequential and a random workload. It gives profile measurements for:
> > +throughput, %CPU rand seeks per second. Not sure if the the CPU numbers
> > +are trustworthy.
> > +"""
> > +
> > +job.run_test('bonnie')
> > diff --git a/client/tests/kvm/autotest_control/dbench.control b/client/tests/kvm/autotest_control/dbench.control
> > new file mode 100644
> > index 0000000..7fb8a37
> > --- /dev/null
> > +++ b/client/tests/kvm/autotest_control/dbench.control
> > @@ -0,0 +1,20 @@
> > +TIME="SHORT"
> > +AUTHOR = "Martin Bligh <mbligh@google.com>"
> > +DOC = """
> > +dbench is one of our standard kernel stress tests.  It produces filesystem
> > +load like netbench originally did, but involves no network system calls.
> > +Its results include throughput rates, which can be used for performance
> > +analysis.
> > +
> > +More information on dbench can be found here:
> > +http://samba.org/ftp/tridge/dbench/README
> > +
> > +Currently it needs to be updated in its configuration. It is a great test for
> > +the higher level I/O systems but barely touches the disk right now.
> > +"""
> > +NAME = 'dbench'
> > +TEST_CLASS = 'kernel'
> > +TEST_CATEGORY = 'Functional'
> > +TEST_TYPE = 'client'
> > +
> > +job.run_test('dbench', seconds=60)
> > diff --git a/client/tests/kvm/autotest_control/sleeptest.control b/client/tests/kvm/autotest_control/sleeptest.control
> > new file mode 100644
> > index 0000000..725ae81
> > --- /dev/null
> > +++ b/client/tests/kvm/autotest_control/sleeptest.control
> > @@ -0,0 +1,15 @@
> > +AUTHOR = "Autotest Team"
> > +NAME = "Sleeptest"
> > +TIME = "SHORT"
> > +TEST_CATEGORY = "Functional"
> > +TEST_CLASS = "General"
> > +TEST_TYPE = "client"
> > +
> > +DOC = """
> > +This test simply sleeps for 1 second by default.  It's a good way to test
> > +profilers and double check that autotest is working.
> > +The seconds argument can also be modified to make the machine sleep for as
> > +long as needed.
> > +"""
> > +
> > +job.run_test('sleeptest', seconds = 1)
> > --
> > 1.6.2.2
> >
> > _______________________________________________
> > Autotest mailing list
> > Autotest@test.kernel.org
> > http://test.kernel.org/cgi-bin/mailman/listinfo/autotest
> >
diff mbox

Patch

diff --git a/client/tests/kvm/autotest_control/bonnie.control b/client/tests/kvm/autotest_control/bonnie.control
new file mode 100644
index 0000000..2717a80
--- /dev/null
+++ b/client/tests/kvm/autotest_control/bonnie.control
@@ -0,0 +1,21 @@ 
+AUTHOR = "Martin Bligh <mbligh@google.com>"
+NAME = "bonnie"
+TIME = "MEDIUM"
+TEST_CLASS = "Kernel"
+TEST_CATEGORY = "Functional"
+TEST_TYPE = "client"
+DOC = """\
+Bonnie is a benchmark which measures the performance of Unix file system
+operations. Bonnie is concerned with identifying bottlenecks; the name is a
+tribute to Bonnie Raitt, who knows how to use one.
+
+For more info, see http://www.textuality.com/bonnie/
+
+This benchmark configuration run generates sustained write traffic
+of 35-50MB/s of .1MB writes to just one disk.  It appears to have a
+sequential and a random workload. It gives profile measurements for:
+throughput, %CPU rand seeks per second. Not sure if the the CPU numbers
+are trustworthy.
+"""
+
+job.run_test('bonnie')
diff --git a/client/tests/kvm/autotest_control/dbench.control b/client/tests/kvm/autotest_control/dbench.control
new file mode 100644
index 0000000..7fb8a37
--- /dev/null
+++ b/client/tests/kvm/autotest_control/dbench.control
@@ -0,0 +1,20 @@ 
+TIME="SHORT"
+AUTHOR = "Martin Bligh <mbligh@google.com>"
+DOC = """
+dbench is one of our standard kernel stress tests.  It produces filesystem
+load like netbench originally did, but involves no network system calls.
+Its results include throughput rates, which can be used for performance
+analysis.
+
+More information on dbench can be found here:
+http://samba.org/ftp/tridge/dbench/README
+
+Currently it needs to be updated in its configuration. It is a great test for
+the higher level I/O systems but barely touches the disk right now.
+"""
+NAME = 'dbench'
+TEST_CLASS = 'kernel'
+TEST_CATEGORY = 'Functional'
+TEST_TYPE = 'client'
+
+job.run_test('dbench', seconds=60)
diff --git a/client/tests/kvm/autotest_control/sleeptest.control b/client/tests/kvm/autotest_control/sleeptest.control
new file mode 100644
index 0000000..725ae81
--- /dev/null
+++ b/client/tests/kvm/autotest_control/sleeptest.control
@@ -0,0 +1,15 @@ 
+AUTHOR = "Autotest Team"
+NAME = "Sleeptest"
+TIME = "SHORT"
+TEST_CATEGORY = "Functional"
+TEST_CLASS = "General"
+TEST_TYPE = "client"
+
+DOC = """
+This test simply sleeps for 1 second by default.  It's a good way to test
+profilers and double check that autotest is working.
+The seconds argument can also be modified to make the machine sleep for as
+long as needed.
+"""
+
+job.run_test('sleeptest', seconds = 1)