| Newsgroups |
gmane.comp.sysutils.pcp |
| Message-ID |
<[email protected]> |
--yabbadabbadoo
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: quoted-printable
[pcp:master] New Comment on Issue #276 Correct mem.vmstat.pgsteal* mem.vmst=
at.pgscan metrics on RHEL7 and Fedora 25 ( https://github.com/performanceco=
pilot/pcp/issues/276 )
By fche ( https://github.com/fche ) :
The way that src/pmdas/linux/proc_vmstat.c is written the metrics are hard=
-coded in C in the vmstat_fields[] array. When it scans /proc/vmstat, it do=
es a linear search against the vmstat_fields[] list to deposit parsed field=
s in the global _proc_vmstat_t structure. It's not particularly efficient, =
nor flexible.
It could instead generate the vmstat pmns entries dynamically from the act=
ual contents of /proc/vmstat, and pass every value out as an int64_t, and d=
itch all the hard-coded proc_vmstat_t etc. machinery.
---
[pcp:master] New Comment on Pull Request #277 perfevent_pmda: Add PMU event=
s exported from kernel ( https://github.com/performancecopilot/pcp/pull/277=
)
By fche ( https://github.com/fche ) :
In addition to any pre-baked /sys/bus fake tarball for qa purposes, please=
consider having another test that actually exercises the code against the =
real running kernel. We've had situations where the pre-baked data gives a =
false assurance, hiding breakage against changes in the real system.
--yabbadabbadoo
Content-Type: text/html; charset="utf-8"
Content-Transfer-Encoding: quoted-printable
[pcp:master] New Comment on Issue <a href=3D"https://github.com/performanc=
ecopilot/pcp/issues/276">#276 Correct mem.vmstat.pgsteal* mem.vmstat.pgscan=
metrics on RHEL7 and Fedora 25</a><br>
By <a href=3D"https://github.com/fche">fche</a>:<p>
<p>The way that src/pmdas/linux/proc_vmstat.c is written the metrics are h=
ard-coded in C in the vmstat_fields[] array. When it scans /proc/vmstat, i=
t does a linear search against the vmstat_fields[] list to deposit parsed f=
ields in the global _proc_vmstat_t structure. It's not particularly effici=
ent, nor flexible.</p>
<p>It could instead generate the vmstat pmns entries dynamically from the =
actual contents of /proc/vmstat, and pass every value out as an int64_t, an=
d ditch all the hard-coded proc_vmstat_t etc. machinery.</p>
<p>
<hr><p>
[pcp:master] New Comment on Pull Request <a href=3D"https://github.com/per=
formancecopilot/pcp/pull/277">#277 perfevent_pmda: Add PMU events exported =
from kernel</a><br>
By <a href=3D"https://github.com/fche">fche</a>:<p>
<p>In addition to any pre-baked /sys/bus fake tarball for qa purposes, ple=
ase consider having another test that actually exercises the code against t=
he real running kernel. We've had situations where the pre-baked data give=
s a false assurance, hiding breakage against changes in the real system.</p=
>
<p>
<div width=3D"1" style=3D"color:white;clear:both">_._,_._,_</div>
<hr>
Groups.io Links:<p>
You receive all messages sent to this group.
<p>
<a target=3D"_blank" href=3D"https://groups.io/g/pcp/message/15237">View/R=
eply Online (#15237)</a> |
<a target=3D"_blank" href=3D"mailto:[email protected]?subject=3DRe:%20%5Bpcp=
%5D%20Updates%20to%20Github%20%23github">Reply To Group</a>
| <a target=3D"_blank" href=3D"mailto:?subject=3DRe:%20%5Bpcp%5D%20Upd=
ates%20to%20Github%20%23github">Reply To Sender</a>
|
<a target=3D"_blank" href=3D"https://groups.io/mt/4406026?uid=3D174580">=
Mute This Topic</a>
=20
| <a href=3D"https://groups.io/g/pcp/post">New Topic</a><br>
<a href=3D"https://groups.io/mk?hashtag=3Dgithub&subid=3D354243">Mute #git=
hub</a>
<p>
<p>pcp mailing list<br />[email protected]<br /><a href=3D"https://groups.io/g=
/pcp/messages" target=3D"_blank">https://groups.io/g/pcp/messages</a></p>
<a href=3D"https://groups.io/g/pcp/editsub?uid=3D174580">Change Your Subsc=
ription</a><br>
<a href=3D"https://groups.io/g/pcp">Group Home</a><br>
<a href=3D"mailto:[email protected]">Contact Group Owner</a><br>
<a href=3D"https://groups.io/static/tos">Terms Of Service</a><br>
<a href=3D"https://groups.io/g/pcp/leave/354243/563757577/xyzzy">Unsubscri=
be From This Group</a><br>
<div width=3D"1" style=3D"color:white;clear:both">_._,_._,_</div>
--yabbadabbadoo--