Hi,

We had a similar problem with our Dell Powerconnect switches no matter what graphing software we used. We had enough traffic that the counters were rotating over the 32bit number or worse staying at the maximum between 5 minute polling intervals.

We increased the frequency of the poller to be every minute and the problem went away. It would be worth trying that to see if the same applies to what you're seeing.

regards,

Paul

On 10/05/2013 07:43, Kristjan Hinn - PERH wrote:

Hei,

Today i hit spikes again. I think i have same problem as http://jira.observium.org/browse/OBSERVIUM-264 as this explains

*From:*observium [mailto:observium-bounces@observium.org] *On Behalf Of *Kristjan Hinn - PERH *Sent:* Wednesday, May 8, 2013 12:07 PM *To:* Observium Network Observation System *Subject:* Re: [Observium] Getting wierd spike graph

As i just started testing observium i only have 10 devices to poll.

*From:*observium [mailto:observium-bounces@observium.org] *On Behalf Of *Adam Armstrong *Sent:* Wednesday, May 8, 2013 11:37 AM *To:* Observium Network Observation System *Subject:* Re: [Observium] Getting wierd spike graph

Why is that strange?

Kristjan Hinn - PERH <Kristjan.Hinn@regionaalhaigla.ee mailto:Kristjan.Hinn@regionaalhaigla.ee> wrote:

Hei,

Looks ok now. It seemed my server could not handle 8 cronjobs. Strange tho, its vm with 2 core and 4GB ram.

*From:*observium [mailto:observium-bounces@observium.org] *On Behalf Of *Kristjan Hinn - PERH *Sent:* Tuesday, May 7, 2013 1:26 PM *To:* Observium Network Observation System *Subject:* Re: [Observium] Getting wierd spike graph

6 switches have one fw and 2 older fw one of them had no spikes.

I cant find any anomalies in network or logs which would cause such spikes. I was using 8 cronjobs for poller. I switched back to one now. No spikes so far. Could that have been the case?

I used cronjobs like mentioned in docs.

*/5 * * * * root /opt/observium/poller.php -i 8 -n 0 >> /dev/null 2>&1

*/5 * * * * root /opt/observium/poller.php -i 8 -n 1 >> /dev/null 2>&1

*/5 * * * * root /opt/observium/poller.php -i 8 -n 2 >> /dev/null 2>&1

*/5 * * * * root /opt/observium/poller.php -i 8 -n 3 >> /dev/null 2>&1

*/5 * * * * root /opt/observium/poller.php -i 8 -n 4 >> /dev/null 2>&1

*/5 * * * * root /opt/observium/poller.php -i 8 -n 5 >> /dev/null 2>&1

*/5 * * * * root /opt/observium/poller.php -i 8 -n 6 >> /dev/null 2>&1

*/5 * * * * root /opt/observium/poller.php -i 8 -n 7 >> /dev/null 2>&1

Now using one poller

*/5 * * * * root /opt/observium/poller.php -h all >> /dev/null 2>&1

*From:*observium [mailto:observium-bounces@observium.org] *On Behalf Of *Adam Armstrong *Sent:* Tuesday, May 7, 2013 11:37 AM *To:* Observium Network Observation System *Subject:* Re: [Observium] Getting wierd spike graph

Its not obvious what the problem is.

The spikes seem to be evenly spaced, which is odd.

Its likely either a bug in the switch SNMP stack which is returning zero or referring to the 32bit counters, or a network congestion issue caused by a crown job.

Do all of the devices have the same software versions?

Adam.

Kristjan Hinn - PERH <Kristjan.Hinn@regionaalhaigla.ee mailto:Kristjan.Hinn@regionaalhaigla.ee> wrote:

Hello,

Im having trouble graphing Avaya ERS5530 switches. Graphs generated have very wierd spikes of ~700 Gbps what is impossible.

I have other switches drawing graphs just fine. And to get this even more wierder one of my eight Avaya ERS5530 is drawing graphs just fine.

Picture of the problematic graph: https://dl.dropboxusercontent.com/u/41978197/observium/xtuum1.JPG

Has anyone got any idea what could case this?

BR,

Kristjan Hinn

---

observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium

Hi,

According to the docs for the powerconnect and the dell "engineers" they should be 64bit values but they just weren't being reported as that for for some reason in the SNMP stack, it's now an internal Dell issue to be fixed.

It is rare but it does happen, decreasing the poll interval may help show if it's an error with the stack reporting or something else :-)

Oh and it's one of the reasons we're going to Force 10 based core switches in the next few months :-)

regards,

Paul

On 10/05/2013 12:52, Adam Armstrong wrote:

If you have devices which only return 32-bit numbers, I would recommend replacing them :)

This issue is mostly likely related to poor connectivity.

Sometimes it's related to the SNMP stack sometimes failing to return 64-bit values, but that's rare.

adam.

On 2013-05-10 09:24, Paul Orrock wrote:

...

Hi,

We had a similar problem with our Dell Powerconnect switches no matter what graphing software we used. We had enough traffic that the counters were rotating over the 32bit number or worse staying at the maximum between 5 minute polling intervals.

We increased the frequency of the poller to be every minute and the problem went away. It would be worth trying that to see if the same applies to what you're seeing.

regards,

Paul

On 10/05/2013 07:43, Kristjan Hinn - PERH wrote: Hei,

Today i hit spikes again. I think i have same problem as http://jira.observium.org/browse/OBSERVIUM-264 as this explains

*From:*observium [mailto:observium-bounces@observium.org] *On Behalf Of *Kristjan Hinn - PERH *Sent:* Wednesday, May 8, 2013 12:07 PM *To:* Observium Network Observation System *Subject:* Re: [Observium] Getting wierd spike graph

As i just started testing observium i only have 10 devices to poll.

*From:*observium [mailto:observium-bounces@observium.org] *On Behalf Of *Adam Armstrong *Sent:* Wednesday, May 8, 2013 11:37 AM *To:* Observium Network Observation System *Subject:* Re: [Observium] Getting wierd spike graph

Why is that strange?

Kristjan Hinn - PERH <Kristjan.Hinn@regionaalhaigla.ee mailto:Kristjan.Hinn@regionaalhaigla.ee> wrote:

Hei,

Looks ok now. It seemed my server could not handle 8 cronjobs. Strange tho, its vm with 2 core and 4GB ram.

*From:*observium [mailto:observium-bounces@observium.org] *On Behalf Of *Kristjan Hinn - PERH *Sent:* Tuesday, May 7, 2013 1:26 PM *To:* Observium Network Observation System *Subject:* Re: [Observium] Getting wierd spike graph

6 switches have one fw and 2 older fw one of them had no spikes.

I cant find any anomalies in network or logs which would cause such spikes. I was using 8 cronjobs for poller. I switched back to one now. No spikes so far. Could that have been the case?

I used cronjobs like mentioned in docs.

*/5 * * * * root /opt/observium/poller.php -i 8 -n 0 >> /dev/null 2>&1

*/5 * * * * root /opt/observium/poller.php -i 8 -n 1 >> /dev/null 2>&1

*/5 * * * * root /opt/observium/poller.php -i 8 -n 2 >> /dev/null 2>&1

*/5 * * * * root /opt/observium/poller.php -i 8 -n 3 >> /dev/null 2>&1

*/5 * * * * root /opt/observium/poller.php -i 8 -n 4 >> /dev/null 2>&1

*/5 * * * * root /opt/observium/poller.php -i 8 -n 5 >> /dev/null 2>&1

*/5 * * * * root /opt/observium/poller.php -i 8 -n 6 >> /dev/null 2>&1

*/5 * * * * root /opt/observium/poller.php -i 8 -n 7 >> /dev/null 2>&1

Now using one poller

*/5 * * * * root /opt/observium/poller.php -h all >> /dev/null 2>&1

*From:*observium [mailto:observium-bounces@observium.org] *On Behalf Of *Adam Armstrong *Sent:* Tuesday, May 7, 2013 11:37 AM *To:* Observium Network Observation System *Subject:* Re: [Observium] Getting wierd spike graph

Its not obvious what the problem is.

The spikes seem to be evenly spaced, which is odd.

Its likely either a bug in the switch SNMP stack which is returning zero or referring to the 32bit counters, or a network congestion issue caused by a crown job.

Do all of the devices have the same software versions?

Adam.

Kristjan Hinn - PERH <Kristjan.Hinn@regionaalhaigla.ee mailto:Kristjan.Hinn@regionaalhaigla.ee> wrote:

Hello,

Im having trouble graphing Avaya ERS5530 switches. Graphs generated have very wierd spikes of ~700 Gbps what is impossible.

I have other switches drawing graphs just fine. And to get this even more wierder one of my eight Avaya ERS5530 is drawing graphs just fine.

Picture of the problematic graph: https://dl.dropboxusercontent.com/u/41978197/observium/xtuum1.JPG

Has anyone got any idea what could case this?

BR,

Kristjan Hinn

---

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

GigE switches without working HC counters are such nonsense :[

(HP 1800 series has it by design, le sigh)

On 10/05/2013 14:06, Paul Orrock wrote:

Hi,

According to the docs for the powerconnect and the dell "engineers" they should be 64bit values but they just weren't being reported as that for for some reason in the SNMP stack, it's now an internal Dell issue to be fixed.

It is rare but it does happen, decreasing the poll interval may help show if it's an error with the stack reporting or something else :-)

Oh and it's one of the reasons we're going to Force 10 based core switches in the next few months :-)

regards,

Paul

On 10/05/2013 12:52, Adam Armstrong wrote:

...

If you have devices which only return 32-bit numbers, I would recommend replacing them :)

This issue is mostly likely related to poor connectivity.

Sometimes it's related to the SNMP stack sometimes failing to return 64-bit values, but that's rare.

adam.

On 2013-05-10 09:24, Paul Orrock wrote:

...

Hi,

We had a similar problem with our Dell Powerconnect switches no matter what graphing software we used. We had enough traffic that the counters were rotating over the 32bit number or worse staying at the maximum between 5 minute polling intervals.

We increased the frequency of the poller to be every minute and the problem went away. It would be worth trying that to see if the same applies to what you're seeing.

regards,

Paul

On 10/05/2013 07:43, Kristjan Hinn - PERH wrote: Hei,

Today i hit spikes again. I think i have same problem as http://jira.observium.org/browse/OBSERVIUM-264 as this explains

*From:*observium [mailto:observium-bounces@observium.org] *On Behalf Of *Kristjan Hinn - PERH *Sent:* Wednesday, May 8, 2013 12:07 PM *To:* Observium Network Observation System *Subject:* Re: [Observium] Getting wierd spike graph

As i just started testing observium i only have 10 devices to poll.

*From:*observium [mailto:observium-bounces@observium.org] *On Behalf Of *Adam Armstrong *Sent:* Wednesday, May 8, 2013 11:37 AM *To:* Observium Network Observation System *Subject:* Re: [Observium] Getting wierd spike graph

Why is that strange?

Kristjan Hinn - PERH <Kristjan.Hinn@regionaalhaigla.ee mailto:Kristjan.Hinn@regionaalhaigla.ee> wrote:

Hei,

Looks ok now. It seemed my server could not handle 8 cronjobs. Strange tho, its vm with 2 core and 4GB ram.

*From:*observium [mailto:observium-bounces@observium.org] *On Behalf Of *Kristjan Hinn - PERH *Sent:* Tuesday, May 7, 2013 1:26 PM *To:* Observium Network Observation System *Subject:* Re: [Observium] Getting wierd spike graph

6 switches have one fw and 2 older fw one of them had no spikes.

I cant find any anomalies in network or logs which would cause such spikes. I was using 8 cronjobs for poller. I switched back to one now. No spikes so far. Could that have been the case?

I used cronjobs like mentioned in docs.

*/5 * * * * root /opt/observium/poller.php -i 8 -n 0 >> /dev/null 2>&1

*/5 * * * * root /opt/observium/poller.php -i 8 -n 1 >> /dev/null 2>&1

*/5 * * * * root /opt/observium/poller.php -i 8 -n 2 >> /dev/null 2>&1

*/5 * * * * root /opt/observium/poller.php -i 8 -n 3 >> /dev/null 2>&1

*/5 * * * * root /opt/observium/poller.php -i 8 -n 4 >> /dev/null 2>&1

*/5 * * * * root /opt/observium/poller.php -i 8 -n 5 >> /dev/null 2>&1

*/5 * * * * root /opt/observium/poller.php -i 8 -n 6 >> /dev/null 2>&1

*/5 * * * * root /opt/observium/poller.php -i 8 -n 7 >> /dev/null 2>&1

Now using one poller

*/5 * * * * root /opt/observium/poller.php -h all >> /dev/null 2>&1

*From:*observium [mailto:observium-bounces@observium.org] *On Behalf Of *Adam Armstrong *Sent:* Tuesday, May 7, 2013 11:37 AM *To:* Observium Network Observation System *Subject:* Re: [Observium] Getting wierd spike graph

Its not obvious what the problem is.

The spikes seem to be evenly spaced, which is odd.

Its likely either a bug in the switch SNMP stack which is returning zero or referring to the 32bit counters, or a network congestion issue caused by a crown job.

Do all of the devices have the same software versions?

Adam.

Kristjan Hinn - PERH <Kristjan.Hinn@regionaalhaigla.ee mailto:Kristjan.Hinn@regionaalhaigla.ee> wrote:

Hello,

Im having trouble graphing Avaya ERS5530 switches. Graphs generated have very wierd spikes of ~700 Gbps what is impossible.

I have other switches drawing graphs just fine. And to get this even more wierder one of my eight Avaya ERS5530 is drawing graphs just fine.

Picture of the problematic graph: https://dl.dropboxusercontent.com/u/41978197/observium/xtuum1.JPG

Has anyone got any idea what could case this?

BR,

Kristjan Hinn

---

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