Needing to reboot the UTM

Hi,

Well the logs are not that informative on this, but you should be able to trace it in the Service Deamon logs.

That being said, uplink errors can be caused by a number of things, usually negotiation with provider router/CPE.

Havent head of any problems yet regarding VMXNET3 and VMWare.

You might want to try to set the wan interface to a forced speed and if possible get the provider to do the same in their end - remember to ask about their MTU values.

If this does not work try to disable automatic monitoring for WAN interface and manually create a new monitoring host

Interfaces ->  Routing- > Uplink Monitoring -> advanced ; create a new Monitoring Host like Google DNS ( create host with IP Address 8.8.8.8 )

If your want interface is having negotiation or MTU issues, the selv monitoring on the interface will report error.

Hope you find the bug :-)

Thanks for the response.

How do I change the interface to force a speed? Is that done in ESXi? I just found in UTM where it gives the interface info, but can't change anything.

If this happened every day, it might be easier/quicker to troubleshoot this :)

Hi,

Check Interfaces & Routing -> Interfaces -> Hardware

If you edit a interface, you will see a checkbox with auto negotiate. If you uncheck this, you should be able to set duplex speed manually.

Come to Think of it, it could be related to the Mac adresse Change on the WAN after system switch. The old Mac could still be in the providers system causing trouble. If you knew the old Mac it could be changed on the WAN interface.

Maybe ask the provider for a line and DHCP  reset

Might be a good idea to debug one system at a time, and not tweak the ESXi until you tried to fix it in the firewall.

When I go to Interfaces & Routing, Interfaces, Hardware, I can only change the HA Link Monitoring, and set a virtual MAC, which I did long ago (Found through trial and error my provider will change my IP address based on the MAC of the WAN interface, and I don't like my IP to change.). There are no more options. It does show more information:

Sorry about that, just learned that this aparently was removed in the virtual image of the utm and is only present on the hardware models now.

Try to check for updated drivers to the esxi.

Looks as If you have to control duplex in vmware.

Refer to vmware kb's like

Make sure you do not have uplink balancing enabled in the utm also.

Well, looks as If trial and error is the way forward then.

Things I might try just to test:

1.) Adding a E1000 adapter to the utm and replace WAN with that ( can be switched directly in esxi cli If you are comftable with that ).

E1000 has its problems, among that is that traffic stops but everything is green. But for the sake of test and driver issues.

2.) place a switch between provider cpe and vmware ( utm ) might solve duplex negation errors, but ofcourse not a long term Solution.

3.) replace netcard to other type in the server If you have something that fits. Ofcourse, just test purpose.

I am more or less out of ideas at this point, Maybe someone else has the answer :-)

Good luck

Hi RChadwick,

I think you will need to change the NIC speed for VM images from the windows network settings, changing it from the Web Admin GUI will not help. You need to change the speed duplex from the Network & settings / Advance / Media Type / NIC Speed. 

Any help?

I changed it from Auto to 1000 full duplex. So far, it's good, but since the problem seems to happen every few weeks on average, it might be a while before I can tell for sure.

Well, changing the speed might have helped (Went a whole month this time), but it happened again today. Restarting the modem, and unplugging/replugging the ethernet WAN cable had no effect. Rebooting the UTM fixed it. Is there by chance a feature that would reboot the UTM automatically if the WAN link shows ERROR?

Check #7.7 in Rulz - you will want to set the modem on 1000/Full fixed also.

Cheers - Bob

Thanks for the response.

The modem doesn't have any adjustments, other than looking through logs and resetting it. I configured ESXi to lock the port at 1000/Full.

A week or two ago I replaced the modem with a newer one. Same manufacturer, same lack of config.

It just happened again a few minutes ago. Is there anything else I can do?

Thanks for the response.

The modem doesn't have any adjustments, other than looking through logs and resetting it. I configured ESXi to lock the port at 1000/Full.

A week or two ago I replaced the modem with a newer one. Same manufacturer, same lack of config.

It just happened again a few minutes ago. Is there anything else I can do?

If it helps anyone with the same problem, I wrote a script running on a VM that checks for Internet, and after 10 minutes reboots the UTM. The script log shows it still happens every few weeks or so, but I haven't noticed it.

^^Can you share?  This would be useful to have.

RebootSophosRouter.zip

I attached my script. It's crude, and I never cleaned it up for public use, but it works. You need to add the IP address of your UTM, and the UTM password in two locations. It pings a few different locations to keep the UTM from rebooting if a single website is down.

^^Thanks for posting.  This runs under windows correct, not as a cron job on the firewall?  I wonder if something like this could be implemented within sophos itself. 

Yes, this runs on a Windows machine, not on the UTM itself.