MikroTik, is there any chance of a test mode, before applying a configuration being incorporated, I have to travel a 30 Mile round trip to site to re-configure because I made a mistake configuring a Powerbox router, apart from that my customer has been off-line all night and this morning and cannot take credit card payments.
On Cisco it’s call ‘reload in’,
Mikrotik have a safe mode https://wiki.mikrotik.com/wiki/Manual:Console#Safe_Mode
That is not a test mode though, that will revert to previous configuration if you screw it up, Ubiquiti kit has a test mode, if you screw up and cannot access the unit, it reverts to previous config and you are back in.
Safe-mode will do the same.
As soon as you lose management connection, it will revert the configuration to the point before safe-mode was engaged.
I think the major difference here (like juniper commit confirmed) is the timeout. So for example if your changing management IP’s you are going to lose connection by virtue of changing management IPs
hence you can’t use safe mode?
Tim
You can also write a small script i think
I have just been to site to have a look at a configuration of a PowerBox I screwed up, and cannot even get to management at site, everything is working except I cannot log on, cannot even see MAC address with Winbox but I can see the CAPS connected to the PowerBox but not the Box itself, while it is working I am not going to factory reset it.
There are very few situations given the flexibility of routeros that safe mode wont work for this purpose. Changing IP/routes? youo can dual stack IP’s and add the routes, safe mode protects perfectly. Firewall changes? Pretty much what safe mode was made for.
I do agree that test change can be beneficial in that you make all the changes, then hit test before it applies it. If you are doing multiple changes in routeros they kind of do apply instantly as you do it (unless you write a script then the delay is minimal) but safe mode should still protect you no matter what, you just might have to think about your order of operations a bit more carefully as to not trigger safe mode revert.
Let me resurrect this topic a bit - I have a following use case:
I need to “revert” the PtP link: as I need to tinker with it a bit, I want the “other” end (to which I have a difficult access) to be a station, and “my” antenna to be a station-bridge.
Now it’s the opposite.
The only mgmt access is trough the radio link itself - so the moment I switch remote “bridge” mode antenna to “station-bridge”, the link goes down. …and safe mode reverts the settings.
Any idea what could I do, or how could this be scripted to have a delay after losing the connection on when the safe mode reverts the changes to the settings?
Look at scheduled scripts. On the remote end, you write a script which sets the items you are going to change back to their current values, and use a scheduler to run it at a certain time in future. Then you make the changes at the remote end first, and then on the local one. If you manage to connect to the remote device after the changes, you remove the scheduler item and the new configuration stays; if you fail to connect, you simply wait until the scheduled time (and revert the changes on the local side in the meantime) and then you’ll be able to try again.
Of course debug every step first:
- create and schedule a test script doing something non-destructive like adding a global variable
- create and run the script that actually restores the current configuration by hand while in safe mode (so that if that script does something wrong, safe mode saves you)