qemu-discuss
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Qemu-discuss] drive cache=writethrough without flush ing to physica


From: Marko Weber | 8000
Subject: Re: [Qemu-discuss] drive cache=writethrough without flush ing to physical disks OR drive cache=none with readonly host cache
Date: Tue, 04 Mar 2014 09:02:13 +0100
User-agent: Roundcube zbfmail Webmail


Hello,

Am 2014-03-03 22:36, schrieb Gionatan Danti:
Il 02/03/2014 22:29, Stanislav German-Evtushenko ha scritto:
Hello Gionatan,

Thank you for the so detailed reply!

    1) simply use the writethrough cache and configure your battery
    backupped RAID card to ignore fsync requests (warning, you really
_need_ a battery backupped RAID card or corruption / data loss will happen in case of a power failure). In this manner, you can use the
    OS-provided read cache, while passing all writes directly to the
RAID card (wich will ignore the fsync requests attached with them).
    The problem is that the writethrough config will impose a big
    overhead on the OS (fsync are not cheap, even when the hardware
    ignore them), resulting in low performance anyway (probably).

It looks as a way to go. I'll check if Dell RAID can be configured to
ignore flushes.

Hi,
with DELL PERC cards (which really are rebranded LSI) w/battery
backup, it should suffice to set the controller's cache to "writeback"
mode. Using the "force writeback" mode disables sync even for RAID
card without battery backup, but this is _strongly_ discouraged.


2) with relatively recent Qemu versions (CentOS 6+), you can use the writeback cache, enable write barriers on the guest OS and configure your battery backupped RAID card to ignore fsync requests (the same
    warning apply: to safely configure your RAID cache as a writeback
    one, you _need_ a battery backupped card). With write barrier
    enabled on the guest side, you are sure that important writes will
    be synched to the RAID cache (as Qemu will pass the relative
    barriers to the underling OS), while the host will cache any read
    requests.


Writeback is an option when you can be sure that all of your systems are
configured properly (with barriers enabled, etc) however this is not
always possible, so I'm trying to find a more general option.

Sure, but both Windows and recent Linux guests enable barriers by
default. An exception is Debian with EXT3 filesystem, where you should
manually enable barriers using the "barrier=1" mount option.


at moment you cant use cache=none when: u use XFS and your XFS formatted with 4k blocks. (mkfs.xfs -b size=4096).
i dont know when this bug will be killed.

marko



Best regards,
Stanislav

Regards.



reply via email to

[Prev in Thread] Current Thread [Next in Thread]