Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 21 Sep 2015 15:11:39 +0100
From:      krad <kraduk@gmail.com>
To:        Quartz <quartz@sneakertech.com>
Cc:        FreeBSD FS <freebsd-fs@freebsd.org>
Subject:   Re: ZFS cpu requirements, with/out compression and/or dedup
Message-ID:  <CALfReydjW6wbSqGCc%2Bvw9-Pb3dgNyOtv7_V5s=xmqa9muNcAQg@mail.gmail.com>
In-Reply-To: <56000CD8.4030208@sneakertech.com>
References:  <CAEW%2BogbPswfOWQzbwNZR5qyMrCEfrcSP4Q7%2By4zuKVVD=KNuUA@mail.gmail.com> <55FD9A2B.8060207@sneakertech.com> <CALfReyc1DcNaRjhhhx%2B4swF2hbfuAd2tWv2xpjWtfqcDoxHUBw@mail.gmail.com> <56000CD8.4030208@sneakertech.com>

next in thread | previous in thread | raw e-mail | index | archive | help
Nope DDT is only used for writes, zfs uses a free block space map, so only
when a block is completely unreferenced will it be written to. The DDT is a
table of blocks and their checksums.

https://blogs.oracle.com/bonwick/en/entry/space_maps

http://www.c0t0d0s0.org/archives/7271-ZFS-Dedup-Internals.html

there are probably better references

On 21 September 2015 at 14:57, Quartz <quartz@sneakertech.com> wrote:

> This is completely untrue,  there performance issues with dedup are
>> limited to writes only, as it needs to check the DDT table for every
>> write to the file system with dedup enabled. Once the data is on the
>> disk there is no overhead, and in many cases a performance boost as less
>> data on the disk means less head movement and its also more likely to be
>> in any available caches. If the write performance does become an issue
>> you can turn it off on that particular file system. This may cause you
>> to no longer have enough capacity on the pool, but then pools are easily
>> extended.
>>
>
> It still needs to keep the tables in memory as long as there's still
> deduped data on disk though, right? Else what keeps track of which blocks
> are used by which files?
>
>
> _______________________________________________
> freebsd-fs@freebsd.org mailing list
> https://lists.freebsd.org/mailman/listinfo/freebsd-fs
> To unsubscribe, send any mail to "freebsd-fs-unsubscribe@freebsd.org"
>



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CALfReydjW6wbSqGCc%2Bvw9-Pb3dgNyOtv7_V5s=xmqa9muNcAQg>