mempool: Remove -persistmempoolv1 option #36294

pull fjahr wants to merge 4 commits into bitcoin:master from fjahr:2026-09-rm-persistmempoolv1 changing 7 files +25 −24

fjahr commented at 7:27 AM on September 18, 2026: contributor

-persistmempoolv1 was added in #28207 as a temporary option to allow downgrading to v26 and earlier, which can't read the XOR'ed mempool.dat. v26 has now been EOL for a while. So any version this option could help downgrade to is unsupported now. It seems ready for removal. I don't think this requires an additional deprecation cycle since the feature is useless for any maintained version and docs have always stated that this was a "temporary option that will be removed in the future".

Reading v1 files is still supported, so upgrading directly from an old version keeps the persisted mempool.

mempool_compatibility.py now uses v28.2 as the old node, the oldest release in get_previous_releases.py that writes the current format.
DrahtBot added the label Mempool on Sep 18, 2026
DrahtBot commented at 7:27 AM on September 18, 2026: contributor


The following sections might be updated with supplementary metadata relevant to reviewers and maintainers.



External sites
- Code coverage & benchmarks: https://corecheck.dev/bitcoin/bitcoin/pulls/36294
- Ralph review <img src=https://review.fish.foo/assets/bitcoin/bitcoin/36294.png alt="Ralph review status"> https://git.fish.foo/bitcoin/bitcoin/pulls/36294#author-ralph


Reviews

See the guideline and AI policy for information on the review process.

Type Reviewers

Concept ACK ismaelsadeeq, musaHaruna

If your review is incorrectly listed, please copy-paste <code></code> into the comment that the bot should ignore.



Conflicts

No conflicts as of last run.
in test/functional/mempool_compatibility.py:31 in 0dc5253539
```
  27 | @@ -28,7 +28,7 @@ def skip_test_if_missing_module(self):
  28 |  
  29 |      def setup_network(self):
  30 |          self.add_nodes(self.num_nodes, versions=[
  31 | -            200100,  # Last release without unbroadcast serialization and without XOR
  32 | +            280200,  # Oldest available release with the current (XOR) mempool.dat format
```
maflcko commented at 8:29 AM on September 18, 2026:

huh? Shouldn't this set -persistmempoolv1=1, so that the reading path is still covered?

Otherwise you are removing coverage (see corecheck)

fjahr commented at 9:05 AM on September 18, 2026:

Right, I got confused here. Added the option and also a comment making the coverage explicit where it happens.

maflcko commented at 11:35 AM on September 18, 2026:

This still drops coverage for LogInfo("Failed to deserialize mempool data on file: %s. Continuing anyway.\n", e.what());, but I guess the coverage there was only implicit (due to missing unbroadcast set), and we don't really care about it anyway?

fjahr commented at 12:41 PM on September 19, 2026:

Right, since we didn't have an explicit check for it, it seems like this coverage was mostly incidental. But some general coverage of a failing import on a malformed mempool file seems like a good idea, so I added that in a separate commit and put it into mempool_persist.py since that seems a more fitting place.
maflcko approved
maflcko commented at 8:30 AM on September 18, 2026: member

lgtm. Seems fine. If anyone really wanted to write this, they can just use a previous release and do a start-stop with `-persistmempoolv1=1

But the test patch looks broken
maflcko added this to the milestone 33.0 on Sep 18, 2026
fjahr force-pushed on Sep 18, 2026
in test/functional/mempool_compatibility.py:32 in 6e0bfbb591
```
  29 |  
  30 |      def setup_network(self):
  31 | -        self.add_nodes(self.num_nodes, versions=[
  32 | -            200100,  # Last release without unbroadcast serialization and without XOR
  33 | +        self.add_nodes(self.num_nodes, extra_args=self.extra_args, versions=[
  34 | +            280200,  # Oldest available release with -persistmempoolv1
```
maflcko commented at 11:35 AM on September 18, 2026:

I think we prefer the last release (aka youngest, not oldest), or the comment should say "Any release with ..."

fjahr commented at 12:23 PM on September 19, 2026:

I changed the version to v31 and updated the comment, though it's going to be stale in a little bit since v32 will still have the option and then that will be the last release with the option. But I am not sure this worth splitting hairs over right now...
maflcko approved
maflcko commented at 11:36 AM on September 18, 2026: member

lgtm

Type	Reviewers
Concept ACK	ismaelsadeeq, musaHaruna

test: Use v31.0 as old node in mempool_compatibility.py

Also removes the -nowallet flag because it is unnecessary since the switch to MiniWallet.

86a31b93d0

mempool: Remove -persistmempoolv1 option d1dfdf8276
doc: Add release note for -persistmempoolv1 removal 4c01625f79
fjahr force-pushed on Sep 19, 2026
test: Check that importmempool fails on a malformed mempool.dat 731b89c09b
DrahtBot added the label CI failed on Sep 19, 2026
DrahtBot removed the label CI failed on Sep 19, 2026
ismaelsadeeq commented at 10:27 AM on September 23, 2026: member

Concept ACK
musaHaruna commented at 9:30 AM on October 6, 2026: contributor

Concept ACK.

I think removing this makes sense since the option was temporary and the versions it helped downgrade to are unsupported now.

Small suggestion for the release note: could we add “Reading existing legacy mempool.dat files remains supported”? This would make it clear that removing the option doesn’t prevent loading the persisted mempool when upgrading from an older version.
fjahr commented at 4:11 PM on October 10, 2026: contributor

Small suggestion for the release note: could we add “Reading existing legacy mempool.dat files remains supported”? This would make it clear that removing the option doesn’t prevent loading the persisted mempool when upgrading from an older version.

Thanks for review! I will add this when I need to push again.

Contributors

Labels

Milestone
33.0

Linked (view graph)

#28207 mempool: Persist with XOR

mempool: Remove -persistmempoolv1 option #36294

External sites

Reviews

Conflicts