-
Notifications
You must be signed in to change notification settings - Fork 674
Duplicate automated IP address assignment #3003
Comments
Hi @avkonst, thank you very much for reporting this, it does indeed look like a bug! |
With:
I am able to reproduce the issue in scenario 1 below:
Output of
This line especially looks suspicious to me:
|
It's because the plugin doesn't have a container or endpoint ID to file the address under, so we pass a |
Hi everyone! I reproduced the scenario described by @avkonst and I obtained the same behavior. I also tried with this:
and
The output of
I also tried with ten containers and they all have the same IP address. It seems that the allocator get stuck when we claim for an IP that is not in the range specified in --ipalloc-range. If we do a GET request to http://127.0.0.1:6784/ip we obtain this:
There is another thing that I would like to understand: in the When you request an address, the So why Thank you a lot! |
Fix use of '_' as container ID in claim outside range
Note the fix is in 1.9.8 just released |
I am executing the following sequence:
What you expected to happen?
The idea is that the allocated IP range is split in half for manual and automated IP address assignment. The first and the third containers should get the IP automatically from the 10.32.0.0/13 range. The second has got IP address explicitly assigned from the reserved range for manual assignment.
What happened?
Duplicate IP address for the second and the third container
How to reproduce it?
The script above
Anything else we need to know?
Environment: virtualbox VM.
Versions:
See with weave 1.9.5 and 1.9.7
docker version 1.13.1
The text was updated successfully, but these errors were encountered: