Date: Sun, 20 Jan 2019 16:41:07 +0100 From: Ben RUBSON <ben.rubson@gmail.com> To: freebsd-fs <freebsd-fs@freebsd.org> Subject: Re: ZFS on Hardware RAID Message-ID: <7C7F0DBB-C751-433A-BCB7-4EBE71BB821D@gmail.com> In-Reply-To: <CAOeNLuqQDJ3O1DgzdkshhiJVXd=6aPCMn86BOSwsJLZRdz21aw@mail.gmail.com> References: <1180280695.63420.1547910313494.JavaMail.zimbra@gausus.net> <92646202.63422.1547910433715.JavaMail.zimbra@gausus.net> <CAOeNLurgn-ep1e=Lq9kgxXK%2By5xqq4ULnudKZAbye59Ys7q96Q@mail.gmail.com> <alpine.BSF.2.21.1901200834470.12592@mail0.time-domain.co.uk> <1691666278.63816.1547976245836.JavaMail.zimbra@gausus.net> <CAOeNLuqQDJ3O1DgzdkshhiJVXd=6aPCMn86BOSwsJLZRdz21aw@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
The only thing I would try with a HW RAID card is to configure one volume per physical disk, so that ZFS can operate as intended. Well, more or less, as there will still be some components in the path which should not be expected, such as the RAID card cache (you have ZFS ZIL & ARC / LARC2, upcoming allocation classes). Thus, an IT-mode HBA, for some dozens of dollars / euros, is really worth the price ! ZFS does not really make sense with only one disk (what's the checksum benefit once it has failed ?), you could go with any other FS... Regarding cache, be sure you can rely on your disks, as some (budget-ones ?) may lie answering data is on disk whereas it has only reached their write cache. You can toggle write cache depending on your disks. camcontrol mode /dev/<disk> -m8 | grep WCE Regarding CPU usage, I'm pretty sure all checksum operations are well hardware-supported, CPU should not be a bottleneck here (disks will first be). Ben
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?7C7F0DBB-C751-433A-BCB7-4EBE71BB821D>