view tests/test-bookmarks-corner-case.t @ 51681:522b4d729e89

mmap: populate the mapping by default

Without pre-population, accessing all data through a mmap can result in many
pagefault, reducing performance significantly. If the mmap is prepopulated, the
performance can no longer get slower than a full read.
(See benchmark number below)

In some cases were very few data is read, prepopulating can be overkill and
slower than populating on access (through page fault). So that behavior can be
controlled when the caller can pre-determine the best behavior.
(See benchmark number below)

In addition, testing with populating in a secondary thread yield great result
combining the best of each approach. This might be implemented in later
changesets.

In all cases, using mmap has a great effect on memory usage when many processes
run in parallel on the same machine.

### Benchmarks

# What did I run

A couple of month back I ran a large benchmark campaign to assess the impact of
various approach for using mmap with the revlog (and other files), it
highlighted a few benchmarks that capture the impact of the changes well. So to
validate this change I checked the following:

- log command displaying various revisions
  (read the changelog index)
- log command displaying the patch of listed revisions
  (read the changelog index, the manifest index and a few files indexes)
- unbundling a few revisions
  (read and write changelog, manifest and few files indexes, and walk the graph
  to update some cache)
- pushing a few revisions
  (read and write changelog, manifest and few files indexes, walk the graph to
  update some cache, performs various accesses locally and remotely during
  discovery)

Benchmarks were run using the default module policy (c+py) and the rust one. No
significant difference were found between the two implementation, so we will
present result using the default policy (unless otherwise specified).

I ran them on a few repositories :

- mercurial: a "public changeset only" copy of mercurial from 2018-08-01 using
  zstd compression and sparse-revlog

- pypy: a copy of pypy from 2018-08-01 using zstd compression and sparse-revlog

- netbeans: a copy of netbeans from 2018-08-01 using zstd compression and
  sparse-revlog

- mozilla-try: a copy of mozilla-try from 2019-02-18 using zstd compression and
  sparse-revlog

- mozilla-try persistent-nodemap: Same as the above but with a persistent
  nodemap. Used for the log --patch benchmark only

# Results

For the smaller repositories (mercurial, pypy), the impact of mmap is almost
imperceptible, other cost dominating the operation. The impact of prepopulating
is undiscernible in the benchmark we ran.

For larger repositories the benchmark support explanation given above:

On netbeans, the log can be about 1% faster without repopulation (for a
difference < 100ms) but unbundle becomes a bit slower, even when small.

### data-env-vars.name                             = netbeans-2018-08-01-zstd-sparse-revlog
  # benchmark.name                                 = hg.command.unbundle
  # benchmark.variants.issue6528                   = disabled
  # benchmark.variants.reuse-external-delta-parent = yes
  # benchmark.variants.revs                        = any-1-extra-rev
  # benchmark.variants.source                      = unbundle
  # benchmark.variants.verbosity                   = quiet
with-populate: 0.240157
no-populate:   0.265087  (+10.38%, +0.02)
  # benchmark.variants.revs                        = any-100-extra-rev
with-populate: 1.459518
no-populate:   1.481290   (+1.49%, +0.02)
 ## benchmark.name                                 = hg.command.push
  # benchmark.variants.explicit-rev                = none
  # benchmark.variants.issue6528                   = disabled
  # benchmark.variants.protocol                    = ssh
  # benchmark.variants.reuse-external-delta-parent = yes
  # benchmark.variants.revs                        = any-1-extra-rev
with-populate: 0.771919
no-populate:   0.792025   (+2.60%, +0.02)
  # benchmark.variants.revs                        = any-100-extra-rev
with-populate: 1.459518
no-populate:   1.481290   (+1.49%, +0.02)


For mozilla-try, the "slow down" from pre-populate for small `hg log` is more
visible, but still small in absolute time. (using rust value for the persistent
nodemap value to be relevant).

### data-env-vars.name            = mozilla-try-2019-02-18-ds2-pnm
  # benchmark.name                = hg.command.log
  # bin-env-vars.hg.flavor        = rust
  # benchmark.variants.patch      = yes
  # benchmark.variants.limit-rev  = 1
with-populate: 0.237813
no-populate:   0.229452   (-3.52%, -0.01)
  # benchmark.variants.limit-rev  = 10
  # benchmark.variants.patch      = yes
with-populate: 1.213578
no-populate:   1.205189
### data-env-vars.name            = mozilla-try-2019-02-18-zstd-sparse-revlog
  # benchmark.variants.limit-rev  = 1000
  # benchmark.variants.patch      = no
  # benchmark.variants.rev        = tip
with-populate: 0.198607
no-populate:   0.195038   (-1.80%, -0.00)

However pre-populating provide a significant boost on more complex operations
like unbundle or push:

### data-env-vars.name                             = mozilla-try-2019-02-18-zstd-sparse-revlog
  # benchmark.name                                 = hg.command.push
  # benchmark.variants.explicit-rev                = none
  # benchmark.variants.issue6528                   = disabled
  # benchmark.variants.protocol                    = ssh
  # benchmark.variants.reuse-external-delta-parent = yes
  # benchmark.variants.revs                        = any-1-extra-rev
with-populate: 4.798632
no-populate:   4.953295   (+3.22%, +0.15)
  # benchmark.variants.revs                        = any-100-extra-rev
with-populate: 4.903618
no-populate:   5.014963   (+2.27%, +0.11)
 ## benchmark.name                                 = hg.command.unbundle
  # benchmark.variants.revs                        = any-1-extra-rev
with-populate: 1.423411
no-populate:   1.585365  (+11.38%, +0.16)
  # benchmark.variants.revs                        = any-100-extra-rev
with-populate: 1.537909
no-populate:   1.688489   (+9.79%, +0.15)

================================
Test corner case around bookmark
================================

This test file is meant to gather test around bookmark that are specific
 enough to not find a place elsewhere.

Test bookmark/changelog race condition
======================================

The data from the bookmark file are filtered to only contains bookmark with
node known to the changelog. If the cache invalidation between these two bits
goes wrong, bookmark can be dropped.

Setup
-----

initial repository setup

  $ hg init bookrace-server
  $ cd bookrace-server
  $ echo a > a
  $ hg add a
  $ hg commit -m root
  $ echo a >> a
  $ hg bookmark book-A
  $ hg commit -m A0
  $ hg up 'desc(root)'
  1 files updated, 0 files merged, 0 files removed, 0 files unresolved
  (leaving bookmark book-A)
  $ echo b > b
  $ hg add b
  $ hg bookmark book-B
  $ hg commit -m B0
  created new head
  $ hg up null
  0 files updated, 0 files merged, 2 files removed, 0 files unresolved
  (leaving bookmark book-B)
  $ hg phase --public --rev 'all()'
  $ hg log -G
  o  changeset:   2:c79985706978
  |  bookmark:    book-B
  |  tag:         tip
  |  parent:      0:6569b5a81c7e
  |  user:        test
  |  date:        Thu Jan 01 00:00:00 1970 +0000
  |  summary:     B0
  |
  | o  changeset:   1:39c28d785860
  |/   bookmark:    book-A
  |    user:        test
  |    date:        Thu Jan 01 00:00:00 1970 +0000
  |    summary:     A0
  |
  o  changeset:   0:6569b5a81c7e
     user:        test
     date:        Thu Jan 01 00:00:00 1970 +0000
     summary:     root
  
  $ hg book
     book-A                    1:39c28d785860
     book-B                    2:c79985706978
  $ cd ..

Add new changeset on each bookmark in distinct clones

  $ hg clone ssh://user@dummy/bookrace-server client-A
  requesting all changes
  adding changesets
  adding manifests
  adding file changes
  added 3 changesets with 3 changes to 2 files (+1 heads)
  new changesets 6569b5a81c7e:c79985706978
  updating to branch default
  2 files updated, 0 files merged, 0 files removed, 0 files unresolved
  $ hg -R client-A update book-A
  1 files updated, 0 files merged, 1 files removed, 0 files unresolved
  (activating bookmark book-A)
  $ echo a >> client-A/a
  $ hg -R client-A commit -m A1
  $ hg clone ssh://user@dummy/bookrace-server client-B
  requesting all changes
  adding changesets
  adding manifests
  adding file changes
  added 3 changesets with 3 changes to 2 files (+1 heads)
  new changesets 6569b5a81c7e:c79985706978
  updating to branch default
  2 files updated, 0 files merged, 0 files removed, 0 files unresolved
  $ hg -R client-B update book-B
  0 files updated, 0 files merged, 0 files removed, 0 files unresolved
  (activating bookmark book-B)
  $ echo b >> client-B/b
  $ hg -R client-B commit -m B1

extension to reproduce the race
-------------------------------

If two process are pushing we want to make sure the following happens:

* process A read changelog
* process B to its full push
* process A read bookmarks
* process A proceed with rest of the push

We build a server side extension for this purpose

  $ cat > bookrace.py << EOF
  > import atexit
  > import os
  > import time
  > from mercurial import bookmarks, error, extensions
  > 
  > def wait(repo):
  >     if not os.path.exists('push-A-started'):
  >         assert repo._currentlock(repo._lockref) is None
  >         assert repo._currentlock(repo._wlockref) is None
  >         repo.ui.status(b'setting raced push up\n')
  >         with open('push-A-started', 'w'):
  >             pass
  >     clock = 300
  >     while not os.path.exists('push-B-done'):
  >         clock -= 1
  >         if clock <= 0:
  >             raise error.Abort("race scenario timed out")
  >         time.sleep(0.1)
  > 
  > def reposetup(ui, repo):
  >     class racedrepo(repo.__class__):
  >         @property
  >         def _bookmarks(self):
  >             wait(self)
  >             return super(racedrepo, self)._bookmarks
  >     repo.__class__ = racedrepo
  > 
  > def e():
  >     with open('push-A-done', 'w'):
  >         pass
  > atexit.register(e)
  > EOF

Actual test
-----------

Start the raced push.

  $ cat >> bookrace-server/.hg/hgrc << EOF
  > [extensions]
  > bookrace=$TESTTMP/bookrace.py
  > EOF
  $ hg push -R client-A -r book-A >push-output.txt 2>&1 &

Wait up to 30 seconds for that push to start.

  $ clock=30
  $ while [ ! -f push-A-started ] && [ $clock -gt 0 ] ; do
  >    clock=`expr $clock - 1`
  >    sleep 1
  > done

Do the other push.

  $ cat >> bookrace-server/.hg/hgrc << EOF
  > [extensions]
  > bookrace=!
  > EOF

  $ hg push -R client-B -r book-B
  pushing to ssh://user@dummy/bookrace-server
  searching for changes
  remote: adding changesets
  remote: adding manifests
  remote: adding file changes
  remote: added 1 changesets with 1 changes to 1 files
  updating bookmark book-B

Signal the raced put that we are done (it waits up to 30 seconds).

  $ touch push-B-done

Wait for the raced push to finish (with the remaning of the initial 30 seconds).

  $ while [ ! -f push-A-done ] && [ $clock -gt 0 ] ; do
  >    clock=`expr $clock - 1`
  >    sleep 1
  > done

Check raced push output.

  $ cat push-output.txt
  pushing to ssh://user@dummy/bookrace-server
  searching for changes
  remote: setting raced push up
  remote has heads on branch 'default' that are not known locally: f26c3b5167d1
  remote: adding changesets
  remote: adding manifests
  remote: adding file changes
  remote: added 1 changesets with 1 changes to 1 files
  updating bookmark book-A

Check result of the push.

  $ hg -R bookrace-server log -G
  o  changeset:   4:9ce3b28c16de
  |  bookmark:    book-A
  |  tag:         tip
  |  parent:      1:39c28d785860
  |  user:        test
  |  date:        Thu Jan 01 00:00:00 1970 +0000
  |  summary:     A1
  |
  | o  changeset:   3:f26c3b5167d1
  | |  bookmark:    book-B
  | |  user:        test
  | |  date:        Thu Jan 01 00:00:00 1970 +0000
  | |  summary:     B1
  | |
  | o  changeset:   2:c79985706978
  | |  parent:      0:6569b5a81c7e
  | |  user:        test
  | |  date:        Thu Jan 01 00:00:00 1970 +0000
  | |  summary:     B0
  | |
  o |  changeset:   1:39c28d785860
  |/   user:        test
  |    date:        Thu Jan 01 00:00:00 1970 +0000
  |    summary:     A0
  |
  o  changeset:   0:6569b5a81c7e
     user:        test
     date:        Thu Jan 01 00:00:00 1970 +0000
     summary:     root
  
  $ hg -R bookrace-server book
     book-A                    4:9ce3b28c16de
     book-B                    3:f26c3b5167d1