git@vger.kernel.org mailing list mirror (one of many)
 help / color / mirror / code / Atom feed
From: "Ævar Arnfjörð Bjarmason" <avarab@gmail.com>
To: Jeff King <peff@peff.net>
Cc: Geert Jansen <gerardu@amazon.com>,
	Junio C Hamano <gitster@pobox.com>,
	"git\@vger.kernel.org" <git@vger.kernel.org>
Subject: Re: [RFC PATCH] index-pack: improve performance on NFS
Date: Fri, 09 Nov 2018 11:11:55 +0100	[thread overview]
Message-ID: <87h8gq5zmc.fsf@evledraar.gmail.com> (raw)
In-Reply-To: <87lg635hz4.fsf@evledraar.gmail.com>


On Thu, Nov 08 2018, Ævar Arnfjörð Bjarmason wrote:

> On Thu, Nov 08 2018, Jeff King wrote:
>
>> On Wed, Nov 07, 2018 at 10:55:24PM +0000, Geert Jansen wrote:
>>
>>> On Mon, Oct 29, 2018 at 07:27:39PM -0400, Jeff King wrote:
>>>
>>> > On Mon, Oct 29, 2018 at 08:36:07PM +0100, Ævar Arnfjörð Bjarmason wrote:
>>> > >  * Re-roll my 4 patch series to include the patch you have in
>>> > >    <20181027093300.GA23974@sigill.intra.peff.net>
>>> >
>>> > I don't think it's quite ready for inclusion as-is. I hope to brush it
>>> > up a bit, but I have quite a backlog of stuff to review, as well.
>>>
>>> We're still quite keen to get this patch included. Is there anything I can do
>>> to help?
>>
>> Yes, testing and review. :)
>>
>> I won't send the series out just yet, as I suspect it could use another
>> read-through on my part. But if you want to peek at it or try some
>> timings, it's available at:
>>
>>   https://github.com/peff/git jk/loose-cache
>
> Just a comment on this from the series:
>
>     Note that it is possible for this to actually be _slower_. We'll do a
>     full readdir() to fill the cache, so if you have a very large number of
>     loose objects and a very small number of lookups, that readdir() may end
>     up more expensive.
>
>     In practice, though, having a large number of loose objects is already a
>     performance problem, which should be fixed by repacking or pruning via
>     git-gc. So on balance, this should be a good tradeoff.
>
> Our biggest repo has a very large number of loose objects at any given
> time, but the vast majority of these are because gc *is* happening very
> frequently and the default expiry policy of 2wks is in effect.
>
> Having a large number of loose objects is not per-se a performance
> problem.
>
> It's a problem if you end up "faulting" to from packs to the loose
> object directory a lot because those objects are still reachable, but if
> they're not reachable that number can grow very large if your ref churn
> is large (so lots of expired loose object production).
>
> Anyway, the series per-se looks good to me. It's particularly nice to
> have some of the ODB cleanup + cleanup in fetch-pack.c
>
> Just wanted to note that in our default (reasonable) config we do
> produce scenarios where this change can still be somewhat pathological,
> so I'm still interested in disabling it entirely given the
> implausibility of what it's guarding against.

Some actual numbers for this for a fairly small repo on NFS, "cold"
cache (hadn't run it in a bit):

    $ time (find objects/?? -type f|wc -l)
    862
    real    0m1.927s

Warm cache:

    $ time (find objects/?? -type f|wc -l)
    872
    real    0m0.151s

Cold cache on a bigger monorepo:

    $ time (find objects/?? -type f|wc -l)
    real    0m4.336s

Warm cache on a bigger monorepo (more ref churn):

    $ time (find objects/?? -type f|wc -l)
    49746
    real    0m1.082s

This on a server where bulk sustained writes of large files are really
fast (up to 1GB/s). It's just these metadata ops that are slow.

I also get cold cache times of up to 6 seconds on:

    time (find $(ls -d objects/??|sort -R) -type f | wc -l)

As opposed max of ~4s without -R, so I suspect there may be some
client/server optimization where things are iterated over in recursive
glob order (pre-fetched?), whereas the cache will try to fill buckets is
it encounters loose objects, so iterate over objects/{00..ff} randomly.

I'm not really leading up to any point here I haven't made already. I
was just curious to try to find some upper bound of overhead if say a
pack with 512 objects is pushed. In that case it's very likely that we
need to fill at least 200/256 buckets.

  reply	other threads:[~2018-11-09 10:12 UTC|newest]

Thread overview: 99+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-10-25 18:38 [RFC PATCH] index-pack: improve performance on NFS Jansen, Geert
2018-10-26  0:21 ` Junio C Hamano
2018-10-26 20:38   ` Ævar Arnfjörð Bjarmason
2018-10-27  7:26     ` Junio C Hamano
2018-10-27  9:33       ` Jeff King
2018-10-27 11:22         ` Ævar Arnfjörð Bjarmason
2018-10-28 22:50           ` [PATCH 0/4] index-pack: optionally turn off SHA-1 collision checking Ævar Arnfjörð Bjarmason
2018-10-30  2:49             ` Geert Jansen
2018-10-30  9:04               ` Junio C Hamano
2018-10-30 18:43             ` [PATCH v2 0/3] index-pack: test updates Ævar Arnfjörð Bjarmason
2018-11-13 20:19               ` [PATCH v3] index-pack: add ability to disable SHA-1 collision check Ævar Arnfjörð Bjarmason
2018-11-14  7:09                 ` Junio C Hamano
2018-11-14 12:40                   ` Ævar Arnfjörð Bjarmason
2018-10-30 18:43             ` [PATCH v2 1/3] pack-objects test: modernize style Ævar Arnfjörð Bjarmason
2018-10-30 18:43             ` [PATCH v2 2/3] pack-objects tests: don't leave test .git corrupt at end Ævar Arnfjörð Bjarmason
2018-10-30 18:43             ` [PATCH v2 3/3] index-pack tests: don't leave test repo dirty " Ævar Arnfjörð Bjarmason
2018-10-28 22:50           ` [PATCH 1/4] pack-objects test: modernize style Ævar Arnfjörð Bjarmason
2018-10-28 22:50           ` [PATCH 2/4] pack-objects tests: don't leave test .git corrupt at end Ævar Arnfjörð Bjarmason
2018-10-28 22:50           ` [PATCH 3/4] index-pack tests: don't leave test repo dirty " Ævar Arnfjörð Bjarmason
2018-10-28 22:50           ` [PATCH 4/4] index-pack: add ability to disable SHA-1 collision check Ævar Arnfjörð Bjarmason
2018-10-29 15:04           ` [RFC PATCH] index-pack: improve performance on NFS Jeff King
2018-10-29 15:09             ` Jeff King
2018-10-29 19:36             ` Ævar Arnfjörð Bjarmason
2018-10-29 23:27               ` Jeff King
2018-11-07 22:55                 ` Geert Jansen
2018-11-08 12:02                   ` Jeff King
2018-11-08 20:58                     ` Geert Jansen
2018-11-08 21:18                       ` Jeff King
2018-11-08 21:55                         ` Geert Jansen
2018-11-08 22:20                     ` Ævar Arnfjörð Bjarmason
2018-11-09 10:11                       ` Ævar Arnfjörð Bjarmason [this message]
2018-11-12 14:31                       ` Jeff King
2018-11-12 14:46                     ` [PATCH 0/9] caching loose objects Jeff King
2018-11-12 14:46                       ` [PATCH 1/9] fsck: do not reuse child_process structs Jeff King
2018-11-12 15:26                         ` Derrick Stolee
2018-11-12 14:47                       ` [PATCH 2/9] submodule--helper: prefer strip_suffix() to ends_with() Jeff King
2018-11-12 18:23                         ` Stefan Beller
2018-11-12 14:48                       ` [PATCH 3/9] rename "alternate_object_database" to "object_directory" Jeff King
2018-11-12 15:30                         ` Derrick Stolee
2018-11-12 15:36                           ` Jeff King
2018-11-12 19:41                             ` Ramsay Jones
2018-11-12 14:48                       ` [PATCH 4/9] sha1_file_name(): overwrite buffer instead of appending Jeff King
2018-11-12 15:32                         ` Derrick Stolee
2018-11-12 14:49                       ` [PATCH 5/9] handle alternates paths the same as the main object dir Jeff King
2018-11-12 15:38                         ` Derrick Stolee
2018-11-12 15:46                           ` Jeff King
2018-11-12 15:50                             ` Derrick Stolee
2018-11-12 14:50                       ` [PATCH 6/9] sha1-file: use an object_directory for " Jeff King
2018-11-12 15:48                         ` Derrick Stolee
2018-11-12 16:09                           ` Jeff King
2018-11-12 19:04                             ` Stefan Beller
2018-11-22 17:42                               ` Jeff King
2018-11-12 18:48                           ` Stefan Beller
2018-11-12 14:50                       ` [PATCH 7/9] object-store: provide helpers for loose_objects_cache Jeff King
2018-11-12 19:24                         ` René Scharfe
2018-11-12 20:16                           ` Jeff King
2018-11-12 14:54                       ` [PATCH 8/9] sha1-file: use loose object cache for quick existence check Jeff King
2018-11-12 16:00                         ` Derrick Stolee
2018-11-12 16:01                         ` Ævar Arnfjörð Bjarmason
2018-11-12 16:21                           ` Jeff King
2018-11-12 22:18                             ` Ævar Arnfjörð Bjarmason
2018-11-12 22:30                               ` Ævar Arnfjörð Bjarmason
2018-11-13 10:02                                 ` Ævar Arnfjörð Bjarmason
2018-11-14 18:21                                   ` René Scharfe
2018-12-02 10:52                                   ` René Scharfe
2018-12-03 22:04                                     ` Jeff King
2018-12-04 21:45                                       ` René Scharfe
2018-12-05  4:46                                         ` Jeff King
2018-12-05  6:02                                           ` René Scharfe
2018-12-05  6:51                                             ` Jeff King
2018-12-05  8:15                                               ` Jeff King
2018-12-05 18:41                                                 ` René Scharfe
2018-12-05 20:17                                                   ` Jeff King
2018-11-12 22:44                             ` Geert Jansen
2018-11-27 20:48                         ` René Scharfe
2018-12-01 19:49                           ` Jeff King
2018-11-12 14:55                       ` [PATCH 9/9] fetch-pack: drop custom loose object cache Jeff King
2018-11-12 19:25                         ` René Scharfe
2018-11-12 19:32                           ` Ævar Arnfjörð Bjarmason
2018-11-12 20:07                             ` Jeff King
2018-11-12 20:13                             ` René Scharfe
2018-11-12 16:02                       ` [PATCH 0/9] caching loose objects Derrick Stolee
2018-11-12 19:10                         ` Stefan Beller
2018-11-09 13:43                   ` [RFC PATCH] index-pack: improve performance on NFS Ævar Arnfjörð Bjarmason
2018-11-09 16:08                     ` Duy Nguyen
2018-11-10 14:04                       ` Ævar Arnfjörð Bjarmason
2018-11-12 14:34                         ` Jeff King
2018-11-12 22:58                     ` Geert Jansen
2018-10-27 14:04         ` Duy Nguyen
2018-10-29 15:18           ` Jeff King
2018-10-29  0:48         ` Junio C Hamano
2018-10-29 15:20           ` Jeff King
2018-10-29 18:43             ` Ævar Arnfjörð Bjarmason
2018-10-29 21:34           ` Geert Jansen
2018-10-29 21:50             ` Jeff King
2018-10-29 22:21               ` Geert Jansen
2018-10-29 22:27             ` Jeff King
2018-10-29 22:35               ` Stefan Beller
2018-10-29 23:29                 ` Jeff King

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

  List information: http://vger.kernel.org/majordomo-info.html

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=87h8gq5zmc.fsf@evledraar.gmail.com \
    --to=avarab@gmail.com \
    --cc=gerardu@amazon.com \
    --cc=git@vger.kernel.org \
    --cc=gitster@pobox.com \
    --cc=peff@peff.net \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
Code repositories for project(s) associated with this public inbox

	https://80x24.org/mirrors/git.git

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).