Mercurial > hg
view tests/test-rebase-newancestor.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.
author | Gregory Szorc <gregory.szorc@gmail.com> |
---|---|
date | Fri, 09 Oct 2015 11:22:01 -0700 |
parents | aa4a1672583e |
children | 42b5193ab5de |
line wrap: on
line source
$ cat >> $HGRCPATH <<EOF > [extensions] > rebase= > > [alias] > tglog = log -G --template "{rev}: '{desc}' {branches}\n" > EOF $ hg init repo $ cd repo $ echo A > a $ echo >> a $ hg ci -Am A adding a $ echo B > a $ echo >> a $ hg ci -m B $ echo C > a $ echo >> a $ hg ci -m C $ hg up -q -C 0 $ echo D >> a $ hg ci -Am AD created new head $ hg tglog @ 3: 'AD' | | o 2: 'C' | | | o 1: 'B' |/ o 0: 'A' $ hg rebase -s 1 -d 3 rebasing 1:0f4f7cb4f549 "B" merging a rebasing 2:30ae917c0e4f "C" merging a saved backup bundle to $TESTTMP/repo/.hg/strip-backup/0f4f7cb4f549-82b3b163-backup.hg (glob) $ hg tglog o 3: 'C' | o 2: 'B' | @ 1: 'AD' | o 0: 'A' $ cd .. Test rebasing of merges with ancestors of the rebase destination - a situation that often happens when trying to recover from repeated merging with a mainline branch. The test case creates a dev branch that contains a couple of merges from the default branch. When rebasing to the default branch, these merges would be merges with ancestors on the same branch. The merges _could_ contain some interesting conflict resolutions or additional changes in the merge commit, but that is mixed up with the actual merge stuff and there is in general no way to separate them. Note: The dev branch contains _no_ changes to f-default. It might be unclear how rebasing of ancestor merges should be handled, but the current behavior with spurious prompts for conflicts in files that didn't change seems very wrong. $ hg init ancestor-merge $ cd ancestor-merge $ touch f-default $ hg ci -Aqm 'default: create f-default' $ hg branch -q dev $ hg ci -qm 'dev: create branch' $ echo stuff > f-dev $ hg ci -Aqm 'dev: f-dev stuff' $ hg up -q default $ echo stuff > f-default $ hg ci -m 'default: f-default stuff' $ hg up -q dev $ hg merge -q default $ hg ci -m 'dev: merge default' $ hg up -q default $ hg rm f-default $ hg ci -m 'default: remove f-default' $ hg up -q dev $ hg merge -q default $ hg ci -m 'dev: merge default' $ hg up -q default $ echo stuff > f-other $ hg ci -Aqm 'default: f-other stuff' $ hg tglog @ 7: 'default: f-other stuff' | | o 6: 'dev: merge default' dev |/| o | 5: 'default: remove f-default' | | | o 4: 'dev: merge default' dev |/| o | 3: 'default: f-default stuff' | | | o 2: 'dev: f-dev stuff' dev | | | o 1: 'dev: create branch' dev |/ o 0: 'default: create f-default' $ hg clone -qU . ../ancestor-merge-2 Full rebase all the way back from branching point: $ hg rebase -r 'only(dev,default)' -d default rebasing 1:1d1a643d390e "dev: create branch" note: rebase of 1:1d1a643d390e created no changes to commit rebasing 2:ec2c14fb2984 "dev: f-dev stuff" rebasing 4:4b019212aaf6 "dev: merge default" remote changed f-default which local deleted use (c)hanged version or leave (d)eleted? c rebasing 6:9455ee510502 "dev: merge default" saved backup bundle to $TESTTMP/ancestor-merge/.hg/strip-backup/1d1a643d390e-43e9e04b-backup.hg (glob) $ hg tglog o 6: 'dev: merge default' | o 5: 'dev: merge default' | o 4: 'dev: f-dev stuff' | @ 3: 'default: f-other stuff' | o 2: 'default: remove f-default' | o 1: 'default: f-default stuff' | o 0: 'default: create f-default' Grafty cherry picking rebasing: $ cd ../ancestor-merge-2 $ hg phase -fdr0: $ hg rebase -r 'children(only(dev,default))' -d default rebasing 2:ec2c14fb2984 "dev: f-dev stuff" rebasing 4:4b019212aaf6 "dev: merge default" remote changed f-default which local deleted use (c)hanged version or leave (d)eleted? c rebasing 6:9455ee510502 "dev: merge default" saved backup bundle to $TESTTMP/ancestor-merge-2/.hg/strip-backup/ec2c14fb2984-62d0b222-backup.hg (glob) $ hg tglog o 7: 'dev: merge default' | o 6: 'dev: merge default' | o 5: 'dev: f-dev stuff' | o 4: 'default: f-other stuff' | o 3: 'default: remove f-default' | o 2: 'default: f-default stuff' | | o 1: 'dev: create branch' dev |/ o 0: 'default: create f-default' $ cd .. Test order of parents of rebased merged with un-rebased changes as p1. $ hg init parentorder $ cd parentorder $ touch f $ hg ci -Aqm common $ touch change $ hg ci -Aqm change $ touch target $ hg ci -Aqm target $ hg up -qr 0 $ touch outside $ hg ci -Aqm outside $ hg merge -qr 1 $ hg ci -m 'merge p1 3=outside p2 1=ancestor' $ hg par changeset: 4:6990226659be tag: tip parent: 3:f59da8fc0fcf parent: 1:dd40c13f7a6f user: test date: Thu Jan 01 00:00:00 1970 +0000 summary: merge p1 3=outside p2 1=ancestor $ hg up -qr 1 $ hg merge -qr 3 $ hg ci -qm 'merge p1 1=ancestor p2 3=outside' $ hg par changeset: 5:a57575f79074 tag: tip parent: 1:dd40c13f7a6f parent: 3:f59da8fc0fcf user: test date: Thu Jan 01 00:00:00 1970 +0000 summary: merge p1 1=ancestor p2 3=outside $ hg tglog @ 5: 'merge p1 1=ancestor p2 3=outside' |\ +---o 4: 'merge p1 3=outside p2 1=ancestor' | |/ | o 3: 'outside' | | +---o 2: 'target' | | o | 1: 'change' |/ o 0: 'common' $ hg rebase -r 4 -d 2 rebasing 4:6990226659be "merge p1 3=outside p2 1=ancestor" saved backup bundle to $TESTTMP/parentorder/.hg/strip-backup/6990226659be-4d67a0d3-backup.hg (glob) $ hg tip changeset: 5:cca50676b1c5 tag: tip parent: 2:a60552eb93fb parent: 3:f59da8fc0fcf user: test date: Thu Jan 01 00:00:00 1970 +0000 summary: merge p1 3=outside p2 1=ancestor $ hg rebase -r 4 -d 2 rebasing 4:a57575f79074 "merge p1 1=ancestor p2 3=outside" saved backup bundle to $TESTTMP/parentorder/.hg/strip-backup/a57575f79074-385426e5-backup.hg (glob) $ hg tip changeset: 5:f9daf77ffe76 tag: tip parent: 2:a60552eb93fb parent: 3:f59da8fc0fcf user: test date: Thu Jan 01 00:00:00 1970 +0000 summary: merge p1 1=ancestor p2 3=outside $ hg tglog @ 5: 'merge p1 1=ancestor p2 3=outside' |\ +---o 4: 'merge p1 3=outside p2 1=ancestor' | |/ | o 3: 'outside' | | o | 2: 'target' | | o | 1: 'change' |/ o 0: 'common' rebase of merge of ancestors $ hg up -qr 2 $ hg merge -qr 3 $ echo 'other change while merging future "rebase ancestors"' > other $ hg ci -Aqm 'merge rebase ancestors' $ hg rebase -d 5 -v rebasing 6:4c5f12f25ebe "merge rebase ancestors" (tip) resolving manifests removing other note: merging f9daf77ffe76+ and 4c5f12f25ebe using bids from ancestors a60552eb93fb and f59da8fc0fcf calculating bids for ancestor a60552eb93fb resolving manifests calculating bids for ancestor f59da8fc0fcf resolving manifests auction for merging merge bids other: consensus for g end of auction getting other committing files: other committing manifest committing changelog rebase merging completed 1 changesets found uncompressed size of bundle content: 193 (changelog) 196 (manifests) 162 other saved backup bundle to $TESTTMP/parentorder/.hg/strip-backup/4c5f12f25ebe-f46990e5-backup.hg (glob) 1 changesets found uncompressed size of bundle content: 252 (changelog) 147 (manifests) 162 other adding branch adding changesets adding manifests adding file changes added 1 changesets with 1 changes to 1 files rebase completed $ hg tglog @ 6: 'merge rebase ancestors' | o 5: 'merge p1 1=ancestor p2 3=outside' |\ +---o 4: 'merge p1 3=outside p2 1=ancestor' | |/ | o 3: 'outside' | | o | 2: 'target' | | o | 1: 'change' |/ o 0: 'common'