lorawan-stack: Upgrade from v3.5.3 to v3.8.4 causes unknown session error

Summary

We have been experiencing an issue with upgrading lorawan stack which is blocking us from upgrading our production lorawan stack to the latest version. When we upgrade our stack in a test environment a large number of devices stop receiving downlinks. …

Steps to Reproduce

Run lorawan-stack v3.5.3
Validate that downlinks work:

$ ttn-lw-cli d downlink replace \
    --application-id myapp \
    --frm-payload 00 \
    --f-port 1 \
    --confirmed \
    --device-id device1

command exits with exit code 0 3. Upgrade to v3.8.4 4. Try send downlink

$ ttn-lw-cli d downlink replace \
    --application-id myapp \
    --frm-payload 00 \
    --f-port 1 \
    --confirmed \
    --device-id device1
error:pkg/networkserver:unknown_session (unknown session)
    correlation_id=596dac8a64e8427ba37b61e122dfed6f

ttn-lw-stack config

What do you see now?

error:pkg/networkserver:unknown_session

What do you want to see instead?

Successful downlink

Environment

…

How do you propose to implement this?

…

How do you propose to test this?

I will be able to text fixes to this using our test environment. We could potentially provide database dumps to help reproduce this issue.

Can you do this yourself and submit a Pull Request?

No. @rvolosatovs

About this issue

Original URL
State: closed
Created 4 years ago
Reactions: 7
Comments: 16 (16 by maintainers)

Most upvoted comments

@KrishnaIyer This is likely the issue. Our start up script runs is-db init, and when it was created it had the same migration behavior as is-db migrate, but it has since changed (https://github.com/TheThingsNetwork/lorawan-stack/commit/f30a16ce0cc42eb47ec9caed9310336e1b00d56c#diff-5e8071a0a861c559b272af7441330c22) so I will update our start up script and run the reproduction steps again to see if we still have the issue.

Thanks!

kurtmc on Jun 25, 2020