From: Milan Broz <mbroz@redhat.com>
To: lvm-devel@redhat.com
Subject: [PATCH 0/5] Fix some scaling problems for large VGs (many LVs)
Date: Mon, 29 Mar 2010 15:12:53 +0200 [thread overview]
Message-ID: <cover.1269866168.git.mbroz@redhat.com> (raw)
This patches fixes some problems in code when processing large VGs.
An example: VG with one PVs and 2000LVs
(I was not able to test more LVs without patches, swapped to death...)
# vgs vg_test
VG #PV #LV #SN Attr VSize VFree
vg_test 1 2000 0 wz--n- 18.53g 10.71g
Before (intentionally do not run any in-kernel operations -t test mode):
# time vgchange -t -a n vg_test
Test mode: Metadata will NOT be updated.
0 logical volume(s) in volume group "vg_test" now active
real 9m31.463s
user 9m29.841s
sys 0m1.251s
After (patches applied):
# time vgchange -t -a n vg_test
Test mode: Metadata will NOT be updated.
0 logical volume(s) in volume group "vg_test" now active
real 1m22.852s
user 1m14.805s
sys 0m7.961s
Memory use (cca, it oscillates according to vg_release)
before:
root 24664 99.9 17.9 355588 350340 pts/0 R+ 14:43 9:34 vgchange -t -a n vg_test
after:
root 31022 98.5 0.2 7736 5084 pts/0 R+ 14:57 1:21 vgchange -t -a n vg_test
...
Milan Broz (5):
Use hash table for quick lv reference when reading metadata.
Remove vg_validate call when parsing cached metadata.
Optimise PV segments search.
Do not traverse PV segment list twice.
Fix all segments memory is allocated from vg private mempool.
lib/cache/lvmcache.c | 3 +-
lib/format_text/import.c | 2 +-
lib/format_text/import_vsn1.c | 42 +++++++++++++++++++++-------
lib/metadata/lv_manip.c | 2 +-
lib/metadata/merge.c | 4 +-
lib/metadata/metadata.c | 12 --------
lib/metadata/metadata.h | 3 --
lib/metadata/pv_alloc.h | 4 ++-
lib/metadata/pv_manip.c | 60 ++++++++++++++++++++++++++++------------
9 files changed, 81 insertions(+), 51 deletions(-)
next reply other threads:[~2010-03-29 13:12 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-03-29 13:12 Milan Broz [this message]
2010-03-29 13:12 ` [PATCH 1/5] Use hash table for quick lv reference when reading metadata Milan Broz
2010-03-29 13:12 ` [PATCH 2/5] Remove vg_validate call when parsing cached metadata Milan Broz
2010-03-31 12:25 ` Zdenek Kabelac
2010-03-29 13:12 ` [PATCH 3/5] Optimise PV segments search Milan Broz
2010-03-29 13:12 ` [PATCH 4/5] Do not traverse PV segment list twice Milan Broz
2010-03-29 13:12 ` [PATCH 5/5] Fix all segments memory is allocated from vg private mempool Milan Broz
2010-03-31 12:35 ` [PATCH 0/5] Fix some scaling problems for large VGs (many LVs) Zdenek Kabelac
2010-03-31 15:06 ` Alasdair G Kergon
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
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=cover.1269866168.git.mbroz@redhat.com \
--to=mbroz@redhat.com \
--cc=lvm-devel@redhat.com \
/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.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.