Date: Tue, 22 Sep 2015 09:42:27 +0200 From: InterNetX - Juergen Gotteswinter <jg@internetx.com> To: freebsd-fs@freebsd.org Subject: Re: [HAST] ZFS, many disks, write order Message-ID: <56010663.3050201@internetx.com> In-Reply-To: <560102BD.6040703@internetx.com> References: <F332F505-F4B3-49D6-8CD7-B53496DCFF3E@gmail.com> <560102BD.6040703@internetx.com>
next in thread | previous in thread | raw e-mail | index | archive | help
my fault, forget my comment about single zil
Am 22.09.2015 um 09:26 schrieb InterNetX - Juergen Gotteswinter:
> One of the great Hast Features is that it Hides Disk Failures from ZFS,
> this has been already discussed on this List a few Months ago.
>
> Single ZIL Flash Drive is also not such a great Idea. I whould suggest
> to start from scratch with "ZFS Best Practices".
>
> Am 21.09.2015 um 23:10 schrieb Ben RUBSON:
>> Hello,
>>
>> I plan to use HAST to synchronize a ZFS pool between 2 servers.
>> The ZFS pool has 3 RAID-Z2 VDEVs (8+2), + 3 spares.
>> +1 mirror for SLOG.
>> +1 mirror for L2ARC.
>> So a total of 37 disks (4TB each).
>> 40Gb/s network bandwidth between the 2 servers.
>>
>> Will I have to define each of the 37 disks/resources in hast.conf ?
>> Stupid question, but will this setup (with so many resources) work ?
>>
>> Will write IOs be ordered on the secondary node in the same order (over all the 37 devices) as they occurred on the primary node ?
>> This of course to have the secondary node consistent, even after a power failure of the primary during a high IO load, leading into an import -F <poolname> on the secondary node.
>>
>> In HAST, each resource seems to be "independent" from the others.
>> In DRBD, as an example, we can put several volumes in a same resource to guarantee write order over all the volumes (disks) of the resource.
>> Example :
>>
>> resource r0 {
>> volume 0 {
>> device /dev/drbd0;
>> disk /dev/c0v0;
>> }
>> volume 1 {
>> device /dev/drbd1;
>> disk /dev/c0v1;
>> }
>> }
>>
>> What about HAST then ?
>> Of course I would have liked to have my 37 disks as volumes in the same resource, as in the example above.
>>
>> Thank you very much for your help !
>>
>> Best regards,
>>
>> Ben
>>
>> _______________________________________________
>> 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"
>>
> _______________________________________________
> 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?56010663.3050201>