From: Zdenek Kabelac <zdenek.kabelac@gmail.com>
To: Mikhail Morfikov <mmorfikov@gmail.com>, linux-lvm@lists.linux.dev
Subject: Re: Should issue_discards be enabled when e2scrub is used on an SSD?
Date: Tue, 2 Jan 2024 12:45:18 +0100 [thread overview]
Message-ID: <d635d089-eb5c-4d51-b41f-f4c61b1aa77c@gmail.com> (raw)
In-Reply-To: <fc9e7a1e-a566-4a5c-8a4a-9e26a5651b65@gmail.com>
Dne 02. 01. 24 v 12:13 Mikhail Morfikov napsal(a):
> On 02/01/2024 11.41, Zdenek Kabelac wrote:
>> Dne 29. 12. 23 v 1:48 Mikhail Morfikov napsal(a):
>>> I've got the following setup: LUKSv2+LVM on an SSD drive, and I
>>> have trim/discad enabled for encrypted container. I also use
>>> e2scrub for Online ext4 Metadata Check which uses a 4 GiB LVM
>>> snapshot. I currently have 10 logical volumes, which are
>>> periodically (once a week) scanned for errors. I've got a question
>>> concerning the *issue_discards* option that can be specified in the
>>> /etc/lvm/lvm.conf file. What actually happens when this option is
>>> enabled and the LVM snapshot is removed after fsck finishes its
>>> job? Will the whole 4 GiB be erased on the flash? Is it safe to
>>> enable this option in such case, or is it better to leave it disabled?
>>
>> Hello
>>
>> issue_discards is relevant only when you *REMOVE* LVs - the free
>> space in VG is after such removal 'discarded' - which is only useful
>> in case you i.e. pay for 'provisioned' space for your VG (or you have
>> some very very very old ssd) - so this will make you VG space
>> consumption smaller - but it will also make restoring of accidentally
>> removed LVs impossible as the recovered LV will be simply just an
>> empty disk space.
>>
>> So for 100% you don't need to enabled it for e2scrub as discards are
>> normally passed through to your LV device as you may easily check in
>> your sysfs dir:
>>
>> /sys/block/XXXX/queue/discard_max_bytes
>>
>> Regards
>>
>> Zdenek
>>
>
> I understand, but the question concerns rather the amount of data written
> to the SSD disk. So I have a 4 GiB LVM snapshot, which will be removed
> after fsck. Will the whole 4 GiB be written to the disk when issue_discards
> is set in the /etc/lvm/lvm.conf file?
The whole size of removed LV will be discard.
It's mostly equivalent of using 'blkdiscard /dev/vg/lv' just before calling
lvremove vg/lv.
lvm2 doesn't interpret in any way what has been used physically used within
the LV.
Regards
Zdenek
next prev parent reply other threads:[~2024-01-02 11:45 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-12-29 0:48 Should issue_discards be enabled when e2scrub is used on an SSD? Mikhail Morfikov
2024-01-02 10:41 ` Zdenek Kabelac
2024-01-02 11:13 ` Mikhail Morfikov
2024-01-02 11:45 ` Zdenek Kabelac [this message]
2024-01-02 12:01 ` Mikhail Morfikov
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=d635d089-eb5c-4d51-b41f-f4c61b1aa77c@gmail.com \
--to=zdenek.kabelac@gmail.com \
--cc=linux-lvm@lists.linux.dev \
--cc=mmorfikov@gmail.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 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).