Hello, On Mon, 6 Mar 2017 16:06:51 +0700 Vy Nguyen Tan wrote: > Hi Jiajia zhong, > > I'm using mixed SSD and HDD on the same node and I did it from url > https://www.sebastien-han.fr/blog/2014/08/25/ceph-mix-sata-and-ssd-within-the-same-box/, > I don't get any problems when run SSD and HDD on the same node. Now I want > to increase Ceph thoughput by increase network into 20Gbps (I want single > network stream get max 20Gbps - test by iperf ). Could you please share > your experience about HA network for Ceph ? What type of bonding do you > have? are you using stackable switches? > This has been discussed here a few times, time to google. ^_^ https://www.spinics.net/lists/ceph-users/msg31660.html In short: 1. Are you _sure_ you actually need the additional bandwidth instead of more PPS/lower latency? 2. Faster links (20Gb/s, 40Gb/s, Infiniband) mean lower latency. 3. Bonding will not improve latency, nor will it give you higher bandwidth for a single stream. 4. Any (ethernet) switches (there are some cheap ones) that support mc-LAG will allow you to do a HA and bonding network. Christian > I very appreciate your help. > > On Mon, Mar 6, 2017 at 11:45 AM, jiajia zhong <zhong2plus at gmail.com> wrote: > > > we are using mixed too, intel PCIE 400G SSD * 8 for metadata pool and tier > > caching pool for our cephfs. > > > > *plus:* > > *'osd crush update on start = false*' as Vladimir replied. > > > > 2017-03-03 20:33 GMT+08:00 ????????????, ???????? <vlad at itgorod.ru>: > > > >> Hi, Matteo! > >> > >> Yes, I'm using mixed cluster in production but it's pretty small at the > >> moment. I've made a smal step by step manual for myself when I did this for > >> the first time and now put it as a gist: https://gist.github.com/vheath > >> en/cf2203aeb53e33e3f80c8c64a02263bc#file-manual-txt. Probably it could > >> be a little bit outdated since it was some time ago. > >> > >> Crush map modifications are going to be persistent in case of reboots > >> and maintenance if you put *'osd crush update on start = false*' to the > >> [osd] section of ceph conf. > >> > >> But I would also recommend to start from this article: > >> https://www.sebastien-han.fr/blog/2014/08/25/ceph- > >> mix-sata-and-ssd-within-the-same-box/ > >> > >> P.S. While I was writing this letter I've seen a letter from Maxime > >> Guyot. Seems that his method is much easier if it leads to the same results. > >> > >> Best regards, > >> Vladimir > >> > >> ? ?????????, > >> ???????????? ???????? > >> ???????? "???? ?????" > >> +7 343 2222192 <+7%20343%20222-21-92> > >> > >> ?????????? ? ??????????? ??????????? > >> IBM, Microsoft, Eset > >> ???????? ???????? "??? ????" > >> ?????????? ??-????? > >> > >> 2017-03-03 16:30 GMT+05:00 Matteo Dacrema <mdacrema at enter.eu>: > >> > >>> Hi all, > >>> > >>> Does anyone run a production cluster with a modified crush map for > >>> create two pools belonging one to HDDs and one to SSDs. > >>> What?s the best method? Modify the crush map via ceph CLI or via text > >>> editor? > >>> Will the modification to the crush map be persistent across reboots and > >>> maintenance operations? > >>> There?s something to consider when doing upgrades or other operations or > >>> different by having ?original? crush map? > >>> > >>> Thank you > >>> Matteo > >>> -------------------------------------------- > >>> This email and any files transmitted with it are confidential and > >>> intended solely for the use of the individual or entity to whom they are > >>> addressed. If you have received this email in error please notify the > >>> system manager. This message contains confidential information and is > >>> intended only for the individual named. If you are not the named addressee > >>> you should not disseminate, distribute or copy this e-mail. Please notify > >>> the sender immediately by e-mail if you have received this e-mail by > >>> mistake and delete this e-mail from your system. If you are not the > >>> intended recipient you are notified that disclosing, copying, distributing > >>> or taking any action in reliance on the contents of this information is > >>> strictly prohibited. > >>> > >>> > >>> _______________________________________________ > >>> ceph-users mailing list > >>> ceph-users at lists.ceph.com > >>> http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com > >>> > >>> > >> > >> _______________________________________________ > >> ceph-users mailing list > >> ceph-users at lists.ceph.com > >> http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com > >> > >> > > > > _______________________________________________ > > ceph-users mailing list > > ceph-users at lists.ceph.com > > http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com > > > > -- Christian Balzer Network/Systems Engineer chibi at gol.com Global OnLine Japan/Rakuten Communications http://www.gol.com/
 |