Oops, one small fix to the numa metrics, on the same branch. I don't
think this is any reason to hold up the 3.1.1 branch, since its already
tagged.
commit 908a60613bace6160ee148284c05f64f2fd4d928
Author: Martin Hicks <mort@xxxxxxxx>
Date: Tue Feb 23 11:10:53 2010 -0500
Fix mem.numa.util.NFS_Unstable metric
The table was being searched for a non-existant key.
Signed-off-by: Martin Hicks <mort@xxxxxxxx>
diff --git a/src/pmdas/linux/pmda.c b/src/pmdas/linux/pmda.c
index 28b8a9c..29fe765 100644
--- a/src/pmdas/linux/pmda.c
+++ b/src/pmdas/linux/pmda.c
@@ -5653,7 +5653,7 @@ linux_fetchCallBack(pmdaMetric *mdesc, unsigned int inst,
pmAtomValue *atom)
break;
case 23: /* mem.numa.util.NFS_Unstable */
- sts = linux_table_lookup("nfs_unstable:",
numa_meminfo.node_info[inst].meminfo,
+ sts = linux_table_lookup("NFS_Unstable:",
numa_meminfo.node_info[inst].meminfo,
&atom->ull);
break;
On Mon, Feb 22, 2010 at 03:56:39PM -0600, Martin Hicks wrote:
>
> new kernels have had additional information put into /proc/meminfo. I
> updated the mem.util metrics to export this additional information in
> Linux.
>
> I did this while I made the NUMA meminfo metrics more consistent with
> the naming of the mem.util metrics...that was a bit sloppy in my first
> pass at the numa meminfo metrics.
>
> git://oss.sgi.com/mort/pcp.git Branch: numa-metrics
>
> commit 83c3d617b53559c497a88726f4619e4b97a46497
> Author: Martin Hicks <mort@xxxxxxxx>
> Date: Mon Feb 22 16:47:13 2010 -0500
>
> Update /proc/meminfo stats on linux
>
> Add some new metrics. Also I made the naming of the NUMA
> statistics more consistent with the naming of those from the
> global meminfo file.
>
> Signed-off-by: Martin Hicks <mort@xxxxxxxx>
>
>
>
|