From: bugzilla-daemon@bugzilla.kernel.org
To: linux-ext4@vger.kernel.org
Subject: [Bug 78651] Write performance of ext4 degrades linearly as volume fills
Date: Sun, 22 Jun 2014 23:18:10 +0000 [thread overview]
Message-ID: <bug-78651-13602-CsIw79ZmUN@https.bugzilla.kernel.org/> (raw)
In-Reply-To: <bug-78651-13602@https.bugzilla.kernel.org/>
https://bugzilla.kernel.org/show_bug.cgi?id=78651
--- Comment #2 from Joseph D. Wagner <joe@josephdwagner.info> ---
I changed the script to run "perf record" on the dump command. I turned
tracing on for all jbd2 and ext4 events. (I had to stop it about half way
through, because the perf.data file was getting too big.) All non-zero results
are below.
Are some of these events supposed to be balanced? For example, there were 8
ext4:ext4_es_shrink_enter, but 0 ext4:ext4_es_shrink_exit. I'm wondering if
some of this code is being called recursively without proper unwinding. It
could explain the recursive loss in performance.
499M ext4:ext4_journal_start: 356 chunks LOST!
252M jbd2:jbd2_handle_stats: 4 chunks LOST!
252M jbd2:jbd2_handle_start: 1 chunks LOST!
252M ext4:ext4_mark_inode_dirty: 3 chunks LOST!
246M ext4:ext4_ext_show_extent: 1 chunks LOST!
246M ext4:ext4_ext_map_blocks_exit: 1 chunks LOST!
246M ext4:ext4_ext_map_blocks_enter: 26 chunks LOST!
246M ext4:ext4_es_lookup_extent_exit
246M ext4:ext4_es_lookup_extent_enter: 1 chunks LOST!
246M ext4:ext4_es_insert_extent
246M ext4:ext4_da_write_end: 6 chunks LOST!
246M ext4:ext4_da_write_begin: 441 chunks LOST!
246M ext4:ext4_da_reserve_space
150 ext4:ext4_drop_inode
77 ext4:ext4_releasepage
17 ext4:ext4_es_cache_extent
8 ext4:ext4_sync_fs
8 ext4:ext4_es_shrink_enter
5 ext4:ext4_es_find_delayed_extent_range_exit
5 ext4:ext4_es_find_delayed_extent_range_enter
3 ext4:ext4_request_inode
3 ext4:ext4_allocate_inode
2 ext4:ext4_ext_load_extent
--
You are receiving this mail because:
You are watching the assignee of the bug.
next prev parent reply other threads:[~2014-06-22 23:18 UTC|newest]
Thread overview: 17+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-06-22 5:44 [Bug 78651] New: Write performance of ext4 degrades linearly as volume fills bugzilla-daemon
2014-06-22 5:45 ` [Bug 78651] " bugzilla-daemon
2014-06-22 23:18 ` bugzilla-daemon [this message]
2014-07-05 17:14 ` bugzilla-daemon
2014-07-06 12:52 ` bugzilla-daemon
2014-07-06 17:57 ` bugzilla-daemon
2014-07-07 6:00 ` bugzilla-daemon
2014-07-07 6:53 ` bugzilla-daemon
2014-07-07 12:29 ` bugzilla-daemon
2014-07-08 13:42 ` bugzilla-daemon
2014-07-08 13:42 ` bugzilla-daemon
2014-07-08 13:43 ` bugzilla-daemon
2014-07-08 13:43 ` bugzilla-daemon
2014-07-08 13:43 ` bugzilla-daemon
2014-07-08 14:23 ` bugzilla-daemon
2014-07-09 3:27 ` bugzilla-daemon
2014-07-09 3:30 ` bugzilla-daemon
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=bug-78651-13602-CsIw79ZmUN@https.bugzilla.kernel.org/ \
--to=bugzilla-daemon@bugzilla.kernel.org \
--cc=linux-ext4@vger.kernel.org \
/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 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 NNTP newsgroup(s).