op-batcher: channel closed due to timeout didn't trigger posting data to the L1/DA layer

[roberto-bayardo]

Bug Description

It's my understanding that when a channel times out, the data from the channel should start getting written to the L1 (or whatever the DA layer is). But I noticed a case where this didn't seem to happen:

The Base devnet has a max channel duration of 300 (== 1 hour) so it should be one hour at most between L1 writes. Despite this our alert fired the other day indicating no data had been posted for over 2 hours. When I looked into this, I see where the channel closed due to duration timeout, one hour since the previous write to L1. But I also see that it changed DA type to "calldata" here, requeued all the blocks, but then no data was posted to the L1 for another hour (when the next channel closed). Image below are the logs immediately following the first channel timeout.

Steps to Reproduce

Very hard to reproduce because you need to force the condition where the optimal DA type changes in between channel timeout.

Expected behavior

I expected that after the channel was closed due to timeout, the data would immediately start getting posted to the L1 instead of it taking another hour.

Environment Information:

Configurations:

Logs:

Additional context

---

⚠️ Notice: Issues that do not include the following sections will be subject to closure:

• Bug Description
• Steps to Reproduce
• Environment Information

Please ensure all required sections are filled out accurately to expedite the debugging process and improve issue resolution efficiency.