Concept ACK, would be good to hear from @eklitzke why the original was closed though

Concept ACK though I'd really like to see benchmarks for changes like this, to know if this has any impact in practice worth making the risk/complication of the change.

Probably worthing having -- rough concept ACK -- however:

• The double-lseek will be slower than fstat() on some networked or detachable filesystems.

• If close-and-uncache is used on just-written-to files, this action is a no-op unless a sync (fdatasync etc.) preceded it.

• Definitely worth benchmarking on multiple filesystems to see the positive or negative value of this change.

Rebased and addressed @laanwj's function rename request.

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

Conflicts

Reviewers, this pull request conflicts with the following ones:

• #26705 (clang-tidy: Fix modernize-use-default-member-init in headers and force to check all headers by hebasto)

If you consider this pull request important, please also help to review the conflicting pull requests. Ideally, start with the one that should be merged first.

Rebased

Adding to the three previous requests for benchmarks:

Do we have any benchmarks showing the practical gains in swiftness? :)

This change seems to slow down IBD.
I benchmarked by syncing a fresh node both on master and this PR (rebased on master) until block height=495000 with -prune=550 and -connect=localhost (syncing from existing local full node).
Master took 5h27mins and this PR took 5h40mins.
#14485 (comment)

This is about 4% difference. Could it be within noise/variance? I mean if you run 3 times master and it takes e.g. 5h27min, 5h03min, 5h42min then you know the result from the PR is within noise. I think that if this makes some impact (positive or negative) it should be observable on smaller tests too, e.g. something that runs 5 or 10 minutes. Then you can repeat a few times and see how much is the variance for both master and the PR.

Is such a workload disk bound? If it is, then does the full node (the source of the sync) compete for disk resources too (that would slow down the syncing node and bring noise)? Was the full node also running the PR?

@vasild you're right, sorry about my last benchmark. I didn't lock the CPU frequency which probably caused a lot of noise.

I re-ran some benchmarks with CPU frequency locked. I rebased the PR locally on master.
The local node used to sync the benchmarked nodes runs on this PR.

Command being timed: "./bitcoind_pr -datadir=/tmp/btc -prune=10000 -stopatheight=400000 -connect=localhost -server=0 -printtoconsole=0"
        User time (seconds): 6765.53
        System time (seconds): 707.99
        Percent of CPU this job got: 137%
        Elapsed (wall clock) time (h:mm:ss or m:ss): 1:30:34
        Average shared text size (kbytes): 0
        Average unshared data size (kbytes): 0
        Average stack size (kbytes): 0
        Average total size (kbytes): 0
        Maximum resident set size (kbytes): 2742480
        Average resident set size (kbytes): 0
        Major (requiring I/O) page faults: 1858
        Minor (reclaiming a frame) page faults: 2802209
        Voluntary context switches: 19966530
        Involuntary context switches: 15899
        Swaps: 0
        File system inputs: 473456
        File system outputs: 333039544
        Socket messages sent: 0
        Socket messages received: 0
        Signals delivered: 0
        Page size (bytes): 4096
        Exit status: 0

Command being timed: "./bitcoin/src/bitcoind -datadir=/tmp/btc -prune=10000 -stopatheight=400000 -connect=localhost -server=0 -printtoconsole=0"
        User time (seconds): 6944.92
        System time (seconds): 707.66
        Percent of CPU this job got: 137%
        Elapsed (wall clock) time (h:mm:ss or m:ss): 1:32:33
        Average shared text size (kbytes): 0
        Average unshared data size (kbytes): 0
        Average stack size (kbytes): 0
        Average total size (kbytes): 0
        Maximum resident set size (kbytes): 2733784
        Average resident set size (kbytes): 0
        Major (requiring I/O) page faults: 2161
        Minor (reclaiming a frame) page faults: 2674124
        Voluntary context switches: 20103862
        Involuntary context switches: 18039
        Swaps: 0
        File system inputs: 536328
        File system outputs: 334433416
        Socket messages sent: 0
        Socket messages received: 0
        Signals delivered: 0
        Page size (bytes): 4096
        Exit status: 0

The results seem to be similar, so ACK 3f2c08b.

🐙 This pull request conflicts with the target branch and needs rebase.

There hasn't been much activity lately and the patch still needs rebase. What is the status here?

• Is it still relevant? ➡️ Please solve the conflicts to make it ready for review and to ensure the CI passes.
• Is it no longer relevant? ➡️ Please close.
• Did the author lose interest or time to work on this? ➡️ Please close it and mark it 'Up for grabs' with the label, so that it can be picked up in the future.

This PR does not seem to have conceptual support. Please leave a comment if you would like this to be reopened.