Netonix Forums

we are having a memory leakage problem. It takes about 9 months and results in the switch staying running but becoming uncontactable, then after a week or so the switch will randomly reboot. It is happening with DC and IDC models (WS-26-500)

What firmware are you running on them? Various memory leak issues have been solved in recent versions.

we are running 1.5.14 on all switches

1.5.16 has additional memory leak fixes. viewtopic.php?f=17&t=240

Unfortunately there is still a memory leak issue with 1.5.16 we believe to be related to discovery (RTGLW is still investigating) , the more devices on the same management IP space will increase this, disabling discovery and rebooting the host in most cases will stop it from happening. you can just toggle discovery when you need to use it, but running 1.5.16 is your best bet until newer stable releases release.

Cdev wrote:Unfortunately there is still a memory leak issue with 1.5.16 we believe to be related to discovery (RTGLW is still investigating) , the more devices on the same management IP space will increase this, disabling discovery and rebooting the host in most cases will stop it from happening. you can just toggle discovery when you need to use it, but running 1.5.16 is your best bet until newer stable releases release.

One of our units must have heard us talking. Looks like we had a switch reboot and most likely was because of a memory event. Looked at our other units and they all seemed to have higher memory usage then I was expecting.

I added a network monitor to track memory usage via SNMP to catch them before rebooting.

JeffreyS wrote:

Cdev wrote:Unfortunately there is still a memory leak issue with 1.5.16 we believe to be related to discovery (RTGLW is still investigating) , the more devices on the same management IP space will increase this, disabling discovery and rebooting the host in most cases will stop it from happening. you can just toggle discovery when you need to use it, but running 1.5.16 is your best bet until newer stable releases release.

One of our units must have heard us talking. Looks like we had a switch reboot and most likely was because of a memory event. Looked at our other units and they all seemed to have higher memory usage then I was expecting.

I added a network monitor to track memory usage via SNMP to catch them before rebooting.

Any updates?

paul.delaney wrote:we are having a memory leakage problem. It takes about 9 months and results in the switch staying running but becoming uncontactable, then after a week or so the switch will randomly reboot. It is happening with DC and IDC models (WS-26-500)

Upgrade to v1.5.21 - BEST VERSION EVER

Only known issue we have to solve yet is mono strand fiber SFP module talking to UBNT mono strand fiber to copper unit.

Memory fixed

SFP issues solved

MANY ISSUES FIXED READ RELEASE NOTES

j2840fl wrote:

JeffreyS wrote:

Cdev wrote:Unfortunately there is still a memory leak issue with 1.5.16 we believe to be related to discovery (RTGLW is still investigating) , the more devices on the same management IP space will increase this, disabling discovery and rebooting the host in most cases will stop it from happening. you can just toggle discovery when you need to use it, but running 1.5.16 is your best bet until newer stable releases release.

One of our units must have heard us talking. Looks like we had a switch reboot and most likely was because of a memory event. Looked at our other units and they all seemed to have higher memory usage then I was expecting.

I added a network monitor to track memory usage via SNMP to catch them before rebooting.

Any updates?

From Oct. 1 to today, all of our switches on 1.5.16 are running low on memory. Only commonality is SNMP being on. 1.5.17 is supposed to have this fixed so I needed upgrade all of them and hopefully they are good.

Many bugs fixed since v1.5.17

USE v1.5.21

If you report issues on any version but latest it will be ignored.