Comprehensive Sub10 support #2469

Merged
merged 2 commits into from Jan 8, 2016

Projects

None yet

8 participants

@thomseddon
Contributor

This adds comprehensive discovery + polling support for Sub10 (recent acquired by Fastback but still distributed under the Sub10 name)

Because of the cost/market positioning (high capacity wireless backhaul) it may not be useful to the user base as a whole in which case I'm happy to just keep & maintain this branch separately :)

@laf laf added Alerting New-Device and removed Alerting labels Nov 20, 2015
@laf
Member
laf commented Nov 20, 2015

Thanks @thomseddon. All looks good to go.

Just one thing :) We've updated the contributors agreement since you first signed it, can you take another look and submit another commit agreeing to the updated version. Just pop a space after your name to edit the file.

Thanks very much.

@laf
Member
laf commented Nov 24, 2015
@f0o
Member
f0o commented Nov 27, 2015

@thomseddon @laf double bump :)

@thomseddon thomseddon commented on the diff Nov 30, 2015
includes/polling/functions.inc.php
@@ -369,8 +369,12 @@ function poll_mib_def($device, $mib_name_table, $mib_subdir, $mib_oids, $mib_gra
$fields = array();
foreach ($oidglist as $fulloid) {
list($splitoid, $splitindex) = explode('.', $fulloid, 2);
- if (is_numeric($snmpdata[$splitindex][$splitoid])) {
- $fields[$oidnamelist[$oid_count]] = $snmpdata[$splitindex][$splitoid];
+ $val = $snmpdata[$splitindex][$splitoid];
+ if (is_numeric($val)) {
+ $rrdupdate .= ':'.$val;
+ }
+ elseif (preg_match("/^\"(.*)\"$/", $val, $number) && is_numeric($number[1])) {
+ $rrdupdate .= ':'.$number[1];
@thomseddon
thomseddon Nov 30, 2015 Contributor

FYI - Sub10 returns some numbers wrapped like so: "20" so this is a workaround - this should maintain backwards compat

@thomseddon
Contributor

Added the extra line - sorry for the delay :)

@laf laf was assigned by f0o Nov 30, 2015
@laf
Member
laf commented Nov 30, 2015

Thanks @thomseddon. All good from our side now expect the agreement, can you add the commit message as described in http://docs.librenms.org/General/Contributing/

Once done, we can merge. Sorry for being picky :/

@laf
Member
laf commented Dec 8, 2015
@laf
Member
laf commented Dec 18, 2015
@f0o
Member
f0o commented Dec 20, 2015
@Rosiak
Contributor
Rosiak commented Jan 5, 2016
@ekoyle
Contributor
ekoyle commented Jan 6, 2016

So it looks like @thomseddon agreed to the new contributor agreement in this PR instead of opening a separate one.

@thomseddon
Contributor

Sorry for epic delay - done in #2717

thomseddon added some commits Sep 11, 2015
@thomseddon thomseddon Add comprehensive Sub10 support 89a76ef
@thomseddon thomseddon Add sub10RadioLclDataRate graph
9593473
@thomseddon
Contributor

Rebased against master also

@Rosiak Rosiak merged commit 89a22ef into librenms:master Jan 8, 2016

2 checks passed

Auto-Deploy Build finished. No test results found.
Details
Scrutinizer 1 new issues
Details
@peppernicus

This commit appears to have broken AirOS polling for wireless stats. A debug poller sample after:

RRD[update /opt/librenms/rrd/silonet-tom/ubnt-airmax-mib.rrd N] ERROR: /opt/librenms/rrd/silonet-tom/ubnt-airmax-mib.rrd: expected timestamp not found in data source from N

whereas before:

RRD[update /opt/librenms/rrd/silonet-tom/ubnt-airmax-mib.rrd N:917:28:750:29:30:-62:34:99:-96:130000000:117000000:1:93:83]

@laf
Member
laf commented Jan 10, 2016

@peppernicus Sorry if this has broken something for you. I actually can't see why this PR would have caused issues.

Can you provide a full output of:

./poller.php -h silonet-tom -d -r

And post the results on dpaste.de or similar.

@peppernicus

Not a big deal - just noticed today that overnight all of the wireless stats graphs for the handful of ubnt radios we run started reporting empty, so it was easy to find the specific commit where it began. Thanks for being part of such a useful tool!

https://dpaste.de/FnVe

@peppernicus

If I had to guess, the changes to includes/polling/functions.inc.php look likely to be responsible.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment