Re: ZFS filesystem - supported ?

From: Mladen Gogala <gogala(dot)mladen(at)gmail(dot)com>
To: pgsql-general(at)lists(dot)postgresql(dot)org
Subject: Re: ZFS filesystem - supported ?
Date: 2021-10-27 00:50:29
Message-ID: 897f283a-5497-734f-b503-406c367dc2d9@gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general


On 10/26/21 20:12, E-BLOKOS wrote:
> RedHat and Oracle are mostly maintaining XFS updates, and I didn't see
> anything saying it's not mainained actively,
> especially when they offering many solutions with XFS as default

Oh, they are maintaining it, all right, but they're not developing it.
XFS is still the file system for rotational disks with plates, reading
heads, tracks and sectors, the type we were taught about in school.
Allocation policy for SSD devices is completely different as are
physical characteristics. Ext4 is being adjusted to ever more popular
NVME devices. XFS is not. In the long run, my money is on Ext4 or its
successors. Here is another useful benchmark:

https://www.percona.com/blog/2012/03/15/ext4-vs-xfs-on-ssd/

This one is a bit old, but it shows clear advantage for Ext4 in async
mode. I maybe wrong. Neither of the two file systems has gained any new
features since 2012. The future may lay in F2FS ("Flash Friendly File
System") which is very new but has a ton of optimizations for SSD
devices. Personally, I usually use XFS for my databases but I am testing
Ext4 with Oracle 21c on Fedora. So far, I don't have any results to
report. The difference is imperceptible. I am primarily an Oracle DBA
and I am testing with Oracle. That doesn't necessarily have to be
pertinent for Postgres.

--
Mladen Gogala
Database Consultant
Tel: (347) 321-1217
https://dbwhisperer.wordpress.com

In response to

Browse pgsql-general by date

  From Date Subject
Next Message Imre Samu 2021-10-27 00:50:44 Re: ZFS filesystem - supported ?
Previous Message Mitar 2021-10-27 00:15:28 Re: Determining if a table really changed in a trigger