ECMP is perfectly fine to use for dual or more wans. Its the least complicated approach. With version 7 firmware it should be the first go to approach. Mangling and PCC come into play for more complex user needs or if the admin has wan throughputs that are wildly dissimilar Accesing services from o...

That is the question... From the beginning with MikroTik, I saw that everyone creates mangle rules to mark connections in dual ISP scenarios when there are services hosted inside the infrastructure for the outside world, and we have 2 ISPs. Coming back to a MikroTik course I took a while ago, the in...

I just try this example and does't seem to work... i just use: "usb2-part1/debian/container1" For some weird reason the alpine image works... :lol: What partitioning is used on the usb2-part1? If it's FAT you'd want to keep the names shorter. And if FAT, maybe use an ext4 partition. it's ...

And make sure you're using the "mount point" shown in "/disk print" (which may not be usb1-part1 in my example). the mount point in my case is usb2-part1, I really would try in the form "root-dir=usb1-part1/some-new-name" from CLI I just try this and does't work... :(

Hmm, if you don't have a preceding slash. Just make sure to pick a new name - I'm not sure what happens if you have had an image file already at the same name. Also keep in mind despite the name "root-dir=", it's actually the name of the container, not a directory to use. So if you keep p...

I try putting a slash behind the path, and on the end, and neither works. In every case it recognise the directory as a container store, but it never works. Now that i think about this.... it makes sense that it doesn't log anything because if this is a path problem, the container never really runs....

i'm gonna try from the terminal... maybe it's winbox? i'm using winbox new version...

What i do is... i select from the dropdown menu the path... and i dont see a slash at the begining of the path... but maybe winbox it's messing the path anyways...

i'll try from the terminal and let you know.

I think i found out a huge bug. Trying in a RB4011, given the extra storage that this router has, i try to download a debian container. When i select a root-dir, it won't start, just like at the begining. Dont know how, but in my 3011 also can't get to work a debian container, but the thing is, with...

i'll try that!

Thanks a lot!

Hope this thread helps more people too.

If you use my example above (https://forum.mikrotik.com/viewtopic.php?t=211025#p1099369), without any custom build, does that work? Hey man! Your example works! Don't know why.... first thought was the nat rule... but i dont see hits in that rule... Things that are different: no root dir setted. no...

i just use this command to build the image and got the same results...

i'll try that. What router do you have? PS: i'll post my export in here for a guy on reddit. ( or for anyone who wants to take a look at it ) # 2024-09-26 10:55:14 by RouterOS 7.16 # software id = U4AN-7YUP # # model = RB3011UiAS # serial number = HF5096ZEDE6 /disk set usb2 media-interface=none medi...

I try deleting the VETH interface that i'm always using to test this, and create another one with another ip address, and same results.

I even try adding this interface to the docker_bridge that i created, and nothing...

It's driving me crazy cause it doesnt even log anything... Setting "logging=yes" isn't enough. You also have to enable the "container" topic: /system/logging/set topics=container action=memory Here's my logging config: https://host.brunolab.com.ar/1.png https://host.brunolab.com...

Does it still not stay started? One note: it's absolutely critical the VETH is created before the container, once the image is first started those setting get persisted to the alpine image. I'm not sure RouterOS updates image if you later change gateway/etc., or at least with alpine. Yes, it still ...

I just do another test.

I install an alpine container that just pings 8.8.8.8.

And i install it on the internal memory.

Same result.

This is the dockerfile:

I forgot...

The entrypoint cant be the problem in this case cause the image itself has the entrypoint configured and..... the famous phrase... it works in my pc

Hi! Thanks for replying! Reading your reply, got me thinking twice in the posibility of doing something wrong in the process, So i re-do everything, and i still the same... I create an httpd container with alpine using buildx for armv7, and imported in a tar file. I try using a pendrive and a samsun...

I manage to get to work a container on a CHR with RouterOS 7.15.3, and replicating the same configuration on my RB3011UiAS does not work at all. Containers wont start. And nothing shows up in the log. (When i hit start, container shows that its running for 1 second and returns to stopped). Yes, i ha...

Hi mate! Were you able to solve your problem? I have the same router, (now in routerOS version 7.15.3), and have the same problem. The container starts, and stops at 1 to 2 seconds. I tryed containers downloaded directly from docker registry and others generated locally and compiled in a tar file. t...

@normis this is very impressive! Congrats on the upgrade! i'm really enjoying the new interface! Feedback: - In IPsec / Indentity, with ROS v 7.14.3, it let me create the identity profile but it forces mode configuration as request only. (router A) - Same happen in ROS 7.15.3 (router B), but when cr...

Also... i just realize that a sync script for an HA cluster just broke completly because of this.

This is so bad...

i'm with ROS 14.3 stable branch and this error persists. I have a full range of scripts that push data to a prometheus pushgateway to audit quality of links and all of them just stop working on update. "failure: there was no content-length or transfer-encoding" Is there any solution to thi...