Juniper MX memory issue
Is anyone else using Observium on the Juniper MX80 platform? I know the MX960 is listed as supported but not the smaller one and we are seeing an odd memory issue that lines up perfectly with when we started polling it with Observium. Attached/linked image shows the last 12 months of memory usage on a MX5. Not sure if it is related to Observium or not so reaching out to see if other are seeing a similar slow memory climb or not.
On 2013-10-12 01:39, Steve Rossen wrote:
Is anyone else using Observium on the Juniper MX80 platform? I know the MX960 is listed as supported but not the smaller one and we are seeing an odd memory issue that lines up perfectly with when we started polling it with Observium. Attached/linked image shows the last 12 months of memory usage on a MX5. Not sure if it is related to Observium or not so reaching out to see if other are seeing a similar slow memory climb or not.
Yes, this is a Juniper bug. It'll keep climbing and then reboot.
One day vendors will get their shit together.
This is /probably/ related to the mac accounting module. The last time I saw this on J kit it was a memory leak in whatever crack-addled code is involved in JUNIPER-MAC-MIB.
adam.
It hit a test box we had. 11.4R8.4 has been safe to date on an MX80 in the lab
Alex
Typographical errors courtesy of iPhone
On Oct 11, 2013, at 21:13, "Adam Armstrong" adama@memetic.org wrote:
On 2013-10-12 01:39, Steve Rossen wrote:
Is anyone else using Observium on the Juniper MX80 platform? I know the MX960 is listed as supported but not the smaller one and we are seeing an odd memory issue that lines up perfectly with when we started polling it with Observium. Attached/linked image shows the last 12 months of memory usage on a MX5. Not sure if it is related to Observium or not so reaching out to see if other are seeing a similar slow memory climb or not. http://i.imgur.com/9PET5L3.png [1]
Yes, this is a Juniper bug. It'll keep climbing and then reboot.
One day vendors will get their shit together.
This is /probably/ related to the mac accounting module. The last time I saw this on J kit it was a memory leak in whatever crack-addled code is involved in JUNIPER-MAC-MIB.
adam. _______________________________________________ observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
To add further details, you should be able to work around this by setting the following two config options in Observium:
$config['poller_modules']['mac-accounting'] = 0; $config['discovery_modules']['mac-accounting'] = 0;
-Dan
On Fri, Oct 11, 2013 at 8:41 PM, Alex Latzko alatzko@servercentral.comwrote:
It hit a test box we had. 11.4R8.4 has been safe to date on an MX80 in the lab
Alex
Typographical errors courtesy of iPhone
On Oct 11, 2013, at 21:13, "Adam Armstrong" adama@memetic.org wrote:
On 2013-10-12 01:39, Steve Rossen wrote:
Is anyone else using Observium on the Juniper MX80 platform? I know the MX960 is listed as supported but not the smaller one and we are seeing an odd memory issue that lines up perfectly with when we started polling it with Observium. Attached/linked image shows the last 12 months of memory usage on a MX5. Not sure if it is related to Observium or not so reaching out to see if other are seeing a similar slow memory climb or not. http://i.imgur.com/9PET5L3.png [1]
Yes, this is a Juniper bug. It'll keep climbing and then reboot.
One day vendors will get their shit together.
This is /probably/ related to the mac accounting module. The last time I
saw this on J kit it was a memory leak in whatever crack-addled code is involved in JUNIPER-MAC-MIB.
adam. _______________________________________________ observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
To add even further detail, only do this if you don't have other non-Juniper or non-bugged devices where you do MAC-accounting; if this is the case, disable the mac accounting modules on the device's config tab.
Tom
On 12/10/2013 9:45, Dan Houtz wrote:
To add further details, you should be able to work around this by setting the following two config options in Observium:
$config['poller_modules']['mac-accounting'] = 0; $config['discovery_modules']['mac-accounting'] = 0;
-Dan
On Fri, Oct 11, 2013 at 8:41 PM, Alex Latzko <alatzko@servercentral.com mailto:alatzko@servercentral.com> wrote:
It hit a test box we had. 11.4R8.4 has been safe to date on an MX80 in the lab Alex Typographical errors courtesy of iPhone On Oct 11, 2013, at 21:13, "Adam Armstrong" <adama@memetic.org <mailto:adama@memetic.org>> wrote: > On 2013-10-12 01:39, Steve Rossen wrote: >> Is anyone else using Observium on the Juniper MX80 platform? I know >> the MX960 is listed as supported but not the smaller one and we are >> seeing an odd memory issue that lines up perfectly with when we >> started polling it with Observium. Attached/linked image shows the >> last 12 months of memory usage on a MX5. Not sure if it is related to >> Observium or not so reaching out to see if other are seeing a similar >> slow memory climb or not. >> http://i.imgur.com/9PET5L3.png [1] > > Yes, this is a Juniper bug. It'll keep climbing and then reboot. > > One day vendors will get their shit together. > > This is /probably/ related to the mac accounting module. The last time I saw this on J kit it was a memory leak in whatever crack-addled code is involved in JUNIPER-MAC-MIB. > > adam. > _______________________________________________ > observium mailing list > observium@observium.org <mailto:observium@observium.org> > http://postman.memetic.org/cgi-bin/mailman/listinfo/observium _______________________________________________ observium mailing list observium@observium.org <mailto:observium@observium.org> http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
Thanks, will try on the version we saw issues on (11.4r4) and the one we are going to (TBD).
On Sat, Oct 12, 2013 at 6:59 AM, Tom Laermans tom.laermans@powersource.cxwrote:
To add even further detail, only do this if you don't have other non-Juniper or non-bugged devices where you do MAC-accounting; if this is the case, disable the mac accounting modules on the device's config tab.
Tom
On 12/10/2013 9:45, Dan Houtz wrote:
To add further details, you should be able to work around this by setting the following two config options in Observium:
$config['poller_modules']['mac-accounting'] = 0; $config['discovery_modules']['mac-accounting'] = 0;
-Dan
On Fri, Oct 11, 2013 at 8:41 PM, Alex Latzko alatzko@servercentral.comwrote:
It hit a test box we had. 11.4R8.4 has been safe to date on an MX80 in the lab
Alex
Typographical errors courtesy of iPhone
On Oct 11, 2013, at 21:13, "Adam Armstrong" adama@memetic.org wrote:
On 2013-10-12 01:39, Steve Rossen wrote:
Is anyone else using Observium on the Juniper MX80 platform? I know the MX960 is listed as supported but not the smaller one and we are seeing an odd memory issue that lines up perfectly with when we started polling it with Observium. Attached/linked image shows the last 12 months of memory usage on a MX5. Not sure if it is related to Observium or not so reaching out to see if other are seeing a similar slow memory climb or not. http://i.imgur.com/9PET5L3.png [1]
Yes, this is a Juniper bug. It'll keep climbing and then reboot.
One day vendors will get their shit together.
This is /probably/ related to the mac accounting module. The last time
I saw this on J kit it was a memory leak in whatever crack-addled code is involved in JUNIPER-MAC-MIB.
adam. _______________________________________________ observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
observium mailing listobservium@observium.orghttp://postman.memetic.org/cgi-bin/mailman/listinfo/observium
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
Hi Steve, I can confirm that there are no issues on 11.4R7.5 either, although I would recommend latest 11.4 at the moment.
Cheers, Florian
On Sat, Oct 12, 2013 at 11:28 PM, Steve Rossen steve.rossen@gmail.com wrote:
Thanks, will try on the version we saw issues on (11.4r4) and the one we are going to (TBD).
On Sat, Oct 12, 2013 at 6:59 AM, Tom Laermans tom.laermans@powersource.cx wrote:
To add even further detail, only do this if you don't have other non-Juniper or non-bugged devices where you do MAC-accounting; if this is the case, disable the mac accounting modules on the device's config tab.
Tom
On 12/10/2013 9:45, Dan Houtz wrote:
To add further details, you should be able to work around this by setting the following two config options in Observium:
$config['poller_modules']['mac-accounting'] = 0; $config['discovery_modules']['mac-accounting'] = 0;
-Dan
On Fri, Oct 11, 2013 at 8:41 PM, Alex Latzko alatzko@servercentral.com wrote:
It hit a test box we had. 11.4R8.4 has been safe to date on an MX80 in the lab
Alex
Typographical errors courtesy of iPhone
On Oct 11, 2013, at 21:13, "Adam Armstrong" adama@memetic.org wrote:
On 2013-10-12 01:39, Steve Rossen wrote:
Is anyone else using Observium on the Juniper MX80 platform? I know the MX960 is listed as supported but not the smaller one and we are seeing an odd memory issue that lines up perfectly with when we started polling it with Observium. Attached/linked image shows the last 12 months of memory usage on a MX5. Not sure if it is related to Observium or not so reaching out to see if other are seeing a similar slow memory climb or not. http://i.imgur.com/9PET5L3.png [1]
Yes, this is a Juniper bug. It'll keep climbing and then reboot.
One day vendors will get their shit together.
This is /probably/ related to the mac accounting module. The last time I saw this on J kit it was a memory leak in whatever crack-addled code is involved in JUNIPER-MAC-MIB.
adam. _______________________________________________ observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
participants (6)
-
Adam Armstrong -
Alex Latzko -
Dan Houtz -
Hibler, Florian -
Steve Rossen -
Tom Laermans