From: Nikolay Borisov <nborisov@suse.com>
To: dsterba@suse.cz, linux-btrfs@vger.kernel.org
Subject: Re: [PATCH 0/5] Misc cleanups in balance code
Date: Fri, 16 Nov 2018 17:36:39 +0200 [thread overview]
Message-ID: <f80bb873-933c-63f7-45a0-57b3ccab46b3@suse.com> (raw)
In-Reply-To: <20181116151806.GZ24115@twin.jikos.cz>
On 16.11.18 г. 17:18 ч., David Sterba wrote:
> On Fri, Oct 26, 2018 at 02:43:16PM +0300, Nikolay Borisov wrote:
>> While investigating the balance hang I came across various inconsistencies in
>> the source. This series aims to fix those.
>>
>> The first patch is (I believe) a fix to a longstanding bug that could cause
>> balance to fail due to ENOSPC. The code no properly ensures that there is
>> at least 1g of unallocated space on every device being balance.
>>
>> Patch 2 makes btrfs_can_relocate a bit more obvious and removes leftovers from
>> previous cleanup
>>
>> Patches 3/4/5 remove some redundant code from various functions.
>>
>> This series has survived multiple xfstest runs.
>>
>> Nikolay Borisov (5):
>> btrfs: Ensure at least 1g is free for balance
>> btrfs: Refactor btrfs_can_relocate
>> btrfs: Remove superfluous check form btrfs_remove_chunk
>> btrfs: Sink find_lock_delalloc_range's 'max_bytes' argument
>> btrfs: Replace BUG_ON with ASSERT in find_lock_delalloc_range
>
> Patches 2-5 on the way to misc-next, thanks. The first one can have user
> visible consequences, so I'd rather first find out why the 1MB was there
> and if it's really a bug and what exactly will change when it's 1G.
I'm fine with that, one thing I don't agree with, though, is the
conclusion that patch 1 has user visible consequence. As a matter of
fact it does not. If btrfs_shrink_device fails with ENOSPC we just
break, i/e we don't try to free balance space for any of the other
devices, but this doesn't stop from balance actually continuing. As it
stands today I think "step one" in __btrfs_balance is more or less
null-op, i.e cycles are wasted since all we do is shrink every device by
1mb, ensuring one more mb is already free. I think this is very
insufficient for a balance operation and when it succeeds during normal
operation it's due to the device already having enough unallocated space
before the balance.
Anyway, I will speak with Chris to try and find out why the code uses 1mb
>
prev parent reply other threads:[~2018-11-16 15:36 UTC|newest]
Thread overview: 25+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-10-26 11:43 [PATCH 0/5] Misc cleanups in balance code Nikolay Borisov
2018-10-26 11:43 ` [PATCH 1/5] btrfs: Ensure at least 1g is free for balance Nikolay Borisov
2018-10-26 12:04 ` Qu Wenruo
2018-10-26 12:08 ` Nikolay Borisov
2018-10-26 12:32 ` Qu Wenruo
2018-10-26 12:09 ` Hans van Kranenburg
2018-10-26 12:16 ` Nikolay Borisov
2018-10-26 12:36 ` Hans van Kranenburg
2018-10-26 11:43 ` [PATCH 2/5] btrfs: Refactor btrfs_can_relocate Nikolay Borisov
2018-10-26 12:35 ` Qu Wenruo
2018-11-17 1:29 ` Anand Jain
2018-12-03 17:25 ` David Sterba
2018-12-04 6:34 ` Nikolay Borisov
2018-12-04 13:07 ` David Sterba
2018-10-26 11:43 ` [PATCH 3/5] btrfs: Remove superfluous check form btrfs_remove_chunk Nikolay Borisov
2018-10-26 12:40 ` Qu Wenruo
2018-11-16 23:57 ` Anand Jain
2018-10-26 11:43 ` [PATCH 4/5] btrfs: Sink find_lock_delalloc_range's 'max_bytes' argument Nikolay Borisov
2018-10-26 12:42 ` Qu Wenruo
2018-11-17 0:53 ` Anand Jain
2018-10-26 11:43 ` [PATCH 5/5] btrfs: Replace BUG_ON with ASSERT in find_lock_delalloc_range Nikolay Borisov
2018-10-26 12:44 ` Qu Wenruo
2018-11-17 1:02 ` Anand Jain
2018-11-16 15:18 ` [PATCH 0/5] Misc cleanups in balance code David Sterba
2018-11-16 15:36 ` Nikolay Borisov [this message]
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=f80bb873-933c-63f7-45a0-57b3ccab46b3@suse.com \
--to=nborisov@suse.com \
--cc=dsterba@suse.cz \
--cc=linux-btrfs@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).