High CPU usage on Juniper EX4300 virtual chassis

We already have different polling models only for Juniper devices because they're so broken. There's not a lot else we can do.

Yes, I do realise that.

I'm sorry to bother you with it, but you probably know, that Juniper won't fix their broken SNMP stack because of customer who's bought 4 low level switches.

You are producing amazing software, that I love and prefer over other alternatives and you'd helped me before in similar situation - update to Junos 17 on EX4500. I thought this time it would be similar resolution. In meantime I fixed the problem by limiting the tree using custom view, but (with all respect) I thought there might be a cooler solution - for example not going through the whole snmp tree every poll, or something like that.

You realise we're not Juniper, right?

The issue is with your switch. Ask Juniper. They're producing switches with broken SNMP stacks.

Hello, still no news about this issue?

Hello, please any news on this issue?

Should I collect and provide some data for you?

attaching  ./poller.php -d -r -h 19 output: ex4300-vc.mngmt.pardubice.cmnet.cz.snmpwalk

trouble with (all) junos device - they have troubles with snmpwalk of whole ifTable oid. And this issue very related to firmware version.

yep, you can make poller debug with this device_id (but this not make sense for issue, since you revert firmware version):

./poller.php -d -r -h 19

mysql> SELECT `device_id` FROM `devices` WHERE `disabled` = 0 AND `hostname` LIKE 'ex4300-vc.mngmt.pardubice.cmnet.cz' AND `poller_id` = '0' ORDER BY `device_id` ASC;
-----------
| device_id |
-----------
|        19 |
-----------
1 row in set (0.00 sec)

Please go to page: device -> performance data
and make screenshot of page.