view tests/test-convert-cvs.t @ 26623:5a95fe44121d

clonebundles: support for seeding clones from pre-generated bundles

Cloning can be an expensive operation for servers because the server
generates a bundle from existing repository data at request time. For
a large repository like mozilla-central, this consumes 4+ minutes
of CPU time on the server. It also results in significant network
utilization. Multiplied by hundreds or even thousands of clients and
the ensuing load can result in difficulties scaling the Mercurial server.

Despite generation of bundles being deterministic until the next
changeset is added, the generation of bundles to service a clone request
is not cached. Each clone thus performs redundant work. This is
wasteful.

This patch introduces the "clonebundles" extension and related
client-side functionality to help alleviate this deficiency. The
client-side feature is behind an experimental flag and is not enabled by
default.

It works as follows:

1) Server operator generates a bundle and makes it available on a
   server (likely HTTP).
2) Server operator defines the URL of a bundle file in a
   .hg/clonebundles.manifest file.
3) Client `hg clone`ing sees the server is advertising bundle URLs.
4) Client fetches and applies the advertised bundle.
5) Client performs equivalent of `hg pull` to fetch changes made since
   the bundle was created.

Essentially, the server performs the expensive work of generating a
bundle once and all subsequent clones fetch a static file from
somewhere. Scaling static file serving is a much more manageable
problem than scaling a Python application like Mercurial. Assuming your
repository grows less than 1% per day, the end result is 99+% of CPU
and network load from clones is eliminated, allowing Mercurial servers
to scale more easily. Serving static files also means data can be
transferred to clients as fast as they can consume it, rather than as
fast as servers can generate it. This makes clones faster.

Mozilla has implemented similar functionality of this patch on
hg.mozilla.org using a custom extension. We are hosting bundle files in
Amazon S3 and CloudFront (a CDN) and have successfully offloaded
>1 TB/day in data transfer from hg.mozilla.org, freeing up significant
bandwidth and CPU resources. The positive impact has been stellar and
I believe it has proved its value to be included in Mercurial core. I
feel it is important for the client-side support to be enabled in core
by default because it means that clients will get faster, more reliable
clones and will enable server operators to reduce load without
requiring any client-side configuration changes (assuming clients are
up to date, of course).

The scope of this feature is narrowly and specifically tailored to
cloning, despite "serve pulls from pre-generated bundles" being a valid
and useful feature. I would eventually like for Mercurial servers to
support transferring *all* repository data via statically hosted files.
You could imagine a server that siphons all pushed data to bundle files
and instructs clients to apply a stream of bundles to reconstruct all
repository data. This feature, while useful and powerful, is
significantly more work to implement because it requires the server
component have awareness of discovery and a mapping of which changesets
are in which files. Full, clone bundles, by contrast, are much simpler.

The wire protocol command is named "clonebundles" instead of something
more generic like "staticbundles" to leave the door open for a new, more
powerful and more generic server-side component with minimal backwards
compatibility implications. The name "bundleclone" is used by Mozilla's
extension and would cause problems since there are subtle differences
in Mozilla's extension.

Mozilla's experience with this idea has taught us that some form of
"content negotiation" is required. Not all clients will support all
bundle formats or even URLs (advanced TLS requirements, etc). To ensure
the highest uptake possible, a server needs to advertise multiple
versions of bundles and clients need to be able to choose the most
appropriate from that list one. The "attributes" in each
server-advertised entry facilitate this filtering and sorting. Their
use will become apparent in subsequent patches.

Initial inspiration and credit for the idea of cloning from static files
belongs to Augie Fackler and his "lookaside clone" extension proof of
concept.

#require cvs

  $ cvscall()
  > {
  >     cvs -f "$@"
  > }
  $ hgcat()
  > {
  >     hg --cwd src-hg cat -r tip "$1"
  > }
  $ echo "[extensions]" >> $HGRCPATH
  $ echo "convert = " >> $HGRCPATH
  $ cat > cvshooks.py <<EOF
  > def cvslog(ui,repo,hooktype,log):
  >     print "%s hook: %d entries"%(hooktype,len(log))
  > 
  > def cvschangesets(ui,repo,hooktype,changesets):
  >     print "%s hook: %d changesets"%(hooktype,len(changesets))
  > EOF
  $ hookpath=`pwd`
  $ cat <<EOF >> $HGRCPATH
  > [hooks]
  > cvslog = python:$hookpath/cvshooks.py:cvslog
  > cvschangesets = python:$hookpath/cvshooks.py:cvschangesets
  > EOF

create cvs repository

  $ mkdir cvsrepo
  $ cd cvsrepo
  $ CVSROOT=`pwd`
  $ export CVSROOT
  $ CVS_OPTIONS=-f
  $ export CVS_OPTIONS
  $ cd ..
  $ rmdir cvsrepo
  $ cvscall -q -d "$CVSROOT" init

create source directory

  $ mkdir src-temp
  $ cd src-temp
  $ echo a > a
  $ mkdir b
  $ cd b
  $ echo c > c
  $ cd ..

import source directory

  $ cvscall -q import -m import src INITIAL start
  N src/a
  N src/b/c
  
  No conflicts created by this import
  
  $ cd ..

checkout source directory

  $ cvscall -q checkout src
  U src/a
  U src/b/c

commit a new revision changing b/c

  $ cd src
  $ sleep 1
  $ echo c >> b/c
  $ cvscall -q commit -mci0 . | grep '<--'
  $TESTTMP/cvsrepo/src/b/c,v  <--  *c (glob)
  $ cd ..

convert fresh repo and also check localtimezone option

NOTE: This doesn't check all time zones -- it merely determines that
the configuration option is taking effect.

An arbitrary (U.S.) time zone is used here.  TZ=US/Hawaii is selected
since it does not use DST (unlike other U.S. time zones) and is always
a fixed difference from UTC.

  $ TZ=US/Hawaii hg convert --config convert.localtimezone=True src src-hg
  initializing destination src-hg repository
  connecting to $TESTTMP/cvsrepo
  scanning source...
  collecting CVS rlog
  5 log entries
  cvslog hook: 5 entries
  creating changesets
  3 changeset entries
  cvschangesets hook: 3 changesets
  sorting...
  converting...
  2 Initial revision
  1 ci0
  0 import
  updating tags
  $ hgcat a
  a
  $ hgcat b/c
  c
  c

convert fresh repo with --filemap

  $ echo include b/c > filemap
  $ hg convert --filemap filemap src src-filemap
  initializing destination src-filemap repository
  connecting to $TESTTMP/cvsrepo
  scanning source...
  collecting CVS rlog
  5 log entries
  cvslog hook: 5 entries
  creating changesets
  3 changeset entries
  cvschangesets hook: 3 changesets
  sorting...
  converting...
  2 Initial revision
  1 ci0
  0 import
  filtering out empty revision
  repository tip rolled back to revision 1 (undo convert)
  updating tags
  $ hgcat b/c
  c
  c
  $ hg -R src-filemap log --template '{rev} {desc} files: {files}\n'
  2 update tags files: .hgtags
  1 ci0 files: b/c
  0 Initial revision files: b/c

convert full repository (issue1649)

  $ cvscall -q -d "$CVSROOT" checkout -d srcfull "." | grep -v CVSROOT
  U srcfull/src/a
  U srcfull/src/b/c
  $ ls srcfull
  CVS
  CVSROOT
  src
  $ hg convert srcfull srcfull-hg \
  >     | grep -v 'log entries' | grep -v 'hook:' \
  >     | grep -v '^[0-3] .*' # filter instable changeset order
  initializing destination srcfull-hg repository
  connecting to $TESTTMP/cvsrepo
  scanning source...
  collecting CVS rlog
  creating changesets
  4 changeset entries
  sorting...
  converting...
  updating tags
  $ hg cat -r tip --cwd srcfull-hg src/a
  a
  $ hg cat -r tip --cwd srcfull-hg src/b/c
  c
  c

commit new file revisions

  $ cd src
  $ echo a >> a
  $ echo c >> b/c
  $ cvscall -q commit -mci1 . | grep '<--'
  $TESTTMP/cvsrepo/src/a,v  <--  a
  $TESTTMP/cvsrepo/src/b/c,v  <--  *c (glob)
  $ cd ..

convert again

  $ TZ=US/Hawaii hg convert --config convert.localtimezone=True src src-hg
  connecting to $TESTTMP/cvsrepo
  scanning source...
  collecting CVS rlog
  7 log entries
  cvslog hook: 7 entries
  creating changesets
  4 changeset entries
  cvschangesets hook: 4 changesets
  sorting...
  converting...
  0 ci1
  $ hgcat a
  a
  a
  $ hgcat b/c
  c
  c
  c

convert again with --filemap

  $ hg convert --filemap filemap src src-filemap
  connecting to $TESTTMP/cvsrepo
  scanning source...
  collecting CVS rlog
  7 log entries
  cvslog hook: 7 entries
  creating changesets
  4 changeset entries
  cvschangesets hook: 4 changesets
  sorting...
  converting...
  0 ci1
  $ hgcat b/c
  c
  c
  c
  $ hg -R src-filemap log --template '{rev} {desc} files: {files}\n'
  3 ci1 files: b/c
  2 update tags files: .hgtags
  1 ci0 files: b/c
  0 Initial revision files: b/c

commit branch

  $ cd src
  $ cvs -q update -r1.1 b/c
  U b/c
  $ cvs -q tag -b branch
  T a
  T b/c
  $ cvs -q update -r branch > /dev/null
  $ sleep 1
  $ echo d >> b/c
  $ cvs -q commit -mci2 . | grep '<--'
  $TESTTMP/cvsrepo/src/b/c,v  <--  *c (glob)
  $ cd ..

convert again

  $ TZ=US/Hawaii hg convert --config convert.localtimezone=True src src-hg
  connecting to $TESTTMP/cvsrepo
  scanning source...
  collecting CVS rlog
  8 log entries
  cvslog hook: 8 entries
  creating changesets
  5 changeset entries
  cvschangesets hook: 5 changesets
  sorting...
  converting...
  0 ci2
  $ hgcat b/c
  c
  d

convert again with --filemap

  $ TZ=US/Hawaii hg convert --config convert.localtimezone=True --filemap filemap src src-filemap
  connecting to $TESTTMP/cvsrepo
  scanning source...
  collecting CVS rlog
  8 log entries
  cvslog hook: 8 entries
  creating changesets
  5 changeset entries
  cvschangesets hook: 5 changesets
  sorting...
  converting...
  0 ci2
  $ hgcat b/c
  c
  d
  $ hg -R src-filemap log --template '{rev} {desc} files: {files}\n'
  4 ci2 files: b/c
  3 ci1 files: b/c
  2 update tags files: .hgtags
  1 ci0 files: b/c
  0 Initial revision files: b/c

commit a new revision with funny log message

  $ cd src
  $ sleep 1
  $ echo e >> a
  $ cvscall -q commit -m'funny
  > ----------------------------
  > log message' . | grep '<--' |\
  >  sed -e 's:.*src/\(.*\),v.*:checking in src/\1,v:g'
  checking in src/a,v

commit new file revisions with some fuzz

  $ sleep 1
  $ echo f >> a
  $ cvscall -q commit -mfuzzy . | grep '<--'
  $TESTTMP/cvsrepo/src/a,v  <--  a
  $ sleep 4 # the two changes will be split if fuzz < 4
  $ echo g >> b/c
  $ cvscall -q commit -mfuzzy . | grep '<--'
  $TESTTMP/cvsrepo/src/b/c,v  <--  *c (glob)
  $ cd ..

convert again

  $ TZ=US/Hawaii hg convert --config convert.cvsps.fuzz=2 --config convert.localtimezone=True src src-hg
  connecting to $TESTTMP/cvsrepo
  scanning source...
  collecting CVS rlog
  11 log entries
  cvslog hook: 11 entries
  creating changesets
  8 changeset entries
  cvschangesets hook: 8 changesets
  sorting...
  converting...
  2 funny
  1 fuzzy
  0 fuzzy
  $ hg -R src-hg log -G --template '{rev} ({branches}) {desc} date: {date|date} files: {files}\n'
  o  8 (branch) fuzzy date: * -1000 files: b/c (glob)
  |
  o  7 (branch) fuzzy date: * -1000 files: a (glob)
  |
  o  6 (branch) funny
  |  ----------------------------
  |  log message date: * -1000 files: a (glob)
  o  5 (branch) ci2 date: * -1000 files: b/c (glob)
  
  o  4 () ci1 date: * -1000 files: a b/c (glob)
  |
  o  3 () update tags date: * +0000 files: .hgtags (glob)
  |
  | o  2 (INITIAL) import date: * -1000 files: (glob)
  | |
  o |  1 () ci0 date: * -1000 files: b/c (glob)
  |/
  o  0 () Initial revision date: * -1000 files: a b/c (glob)
  

testing debugcvsps

  $ cd src
  $ hg debugcvsps --fuzz=2 -x >/dev/null

commit a new revision changing a and removing b/c

  $ cvscall -q update -A
  U a
  U b/c
  $ sleep 1
  $ echo h >> a
  $ cvscall -Q remove -f b/c
  $ cvscall -q commit -mci | grep '<--'
  $TESTTMP/cvsrepo/src/a,v  <--  a
  $TESTTMP/cvsrepo/src/b/c,v  <--  *c (glob)

update and verify the cvsps cache

  $ hg debugcvsps --fuzz=2 -u
  collecting CVS rlog
  13 log entries
  cvslog hook: 13 entries
  creating changesets
  11 changeset entries
  cvschangesets hook: 11 changesets
  ---------------------
  PatchSet 1 
  Date: * (glob)
  Author: * (glob)
  Branch: HEAD
  Tag: (none) 
  Branchpoints: INITIAL 
  Log:
  Initial revision
  
  Members: 
  	a:INITIAL->1.1 
  
  ---------------------
  PatchSet 2 
  Date: * (glob)
  Author: * (glob)
  Branch: HEAD
  Tag: (none) 
  Branchpoints: INITIAL, branch 
  Log:
  Initial revision
  
  Members: 
  	b/c:INITIAL->1.1 
  
  ---------------------
  PatchSet 3 
  Date: * (glob)
  Author: * (glob)
  Branch: INITIAL
  Tag: start 
  Log:
  import
  
  Members: 
  	a:1.1->1.1.1.1 
  	b/c:1.1->1.1.1.1 
  
  ---------------------
  PatchSet 4 
  Date: * (glob)
  Author: * (glob)
  Branch: HEAD
  Tag: (none) 
  Log:
  ci0
  
  Members: 
  	b/c:1.1->1.2 
  
  ---------------------
  PatchSet 5 
  Date: * (glob)
  Author: * (glob)
  Branch: HEAD
  Tag: (none) 
  Branchpoints: branch 
  Log:
  ci1
  
  Members: 
  	a:1.1->1.2 
  
  ---------------------
  PatchSet 6 
  Date: * (glob)
  Author: * (glob)
  Branch: HEAD
  Tag: (none) 
  Log:
  ci1
  
  Members: 
  	b/c:1.2->1.3 
  
  ---------------------
  PatchSet 7 
  Date: * (glob)
  Author: * (glob)
  Branch: branch
  Tag: (none) 
  Log:
  ci2
  
  Members: 
  	b/c:1.1->1.1.2.1 
  
  ---------------------
  PatchSet 8 
  Date: * (glob)
  Author: * (glob)
  Branch: branch
  Tag: (none) 
  Log:
  funny
  ----------------------------
  log message
  
  Members: 
  	a:1.2->1.2.2.1 
  
  ---------------------
  PatchSet 9 
  Date: * (glob)
  Author: * (glob)
  Branch: branch
  Tag: (none) 
  Log:
  fuzzy
  
  Members: 
  	a:1.2.2.1->1.2.2.2 
  
  ---------------------
  PatchSet 10 
  Date: * (glob)
  Author: * (glob)
  Branch: branch
  Tag: (none) 
  Log:
  fuzzy
  
  Members: 
  	b/c:1.1.2.1->1.1.2.2 
  
  ---------------------
  PatchSet 11 
  Date: * (glob)
  Author: * (glob)
  Branch: HEAD
  Tag: (none) 
  Log:
  ci
  
  Members: 
  	a:1.2->1.3 
  	b/c:1.3->1.4(DEAD) 
  

  $ cd ..