If deep-scrubbing is causing a performance problem, you're going to be very unhappy during recovery. I think you'd be better off improving performance so that you can handle normal deep-scrubbing. If you still want to proceed, I wouldn't set the nodeep-scrub flag. I'd adjust the scrubbing variables: osd scrub load threshold osd scrub min interval osd scrub max interval osd deep scrub interval (See https://ceph.com/docs/master/rados/configuration/osd-config-ref/#scrubbing) If you increase the deep scrub interval to something like a month, then run a weekly cron, that should solve your problem without the warning. If the cron fails, deep-scrubbing will still happen once a month. Deep-scrubbing manually is relatively easy, run ceph pg deep-scrub <pg_id>. You can get a list of the pg_id's from ceph pg dump. If you're doing that in a script, you might prefer ceph pg dump --format=json. The pg dump contains last_scrub, scrub_stamp, last_deep_scrub, and deep_scrub_stamp for each PG. You might want to sort the PGs by last_deep_scrub, and do the oldest one first. There's a lot of information here that your script can use to make this manually scrubbing work better for you. On 5/20/14 02:29 , Ta Ba Tuan wrote: > Dear Yang, > > I planed set nodeep-scrub at nigh daily by crontab. > and with error "HEALTH_WARN nodeep-scrub flag(s) set". I only > concentrate messages from the monitoring tool (vd: nagios) => and I > re-writed nagios'checkscript to with message "HEALTH_WARN nodeep-scrub > flag(s) set" returns code = 0. > > > > > On 05/20/2014 10:47 AM, Jianing Yang wrote: >> I found that deep scrub has a significant impact on my cluster. I've >> used "ceph osd set nodeep-scrub" disable it. But I got an error >> "HEALTH_WARN nodeep-scrub flag(s) set". What is the proper way to >> disable deep scrub? and how can I run it manually? >> >> -- >> _____________________________________ >> / Install 'denyhosts' to help protect \ >> | against brute force SSH attacks, | >> \ auto-blocking multiple attempts. / >> ------------------------------------- >> \ >> \ >> \ >> .- <O> -. .-====-. ,-------. .-=<>=-. >> /_-\'''/-_\ / / '' \ \ |,-----.| /__----__\ >> |/ o) (o \| | | ')(' | | /,'-----'.\ |/ (')(') \| >> \ ._. / \ \ / / {_/(') (')\_} \ __ / >> ,>-_,,,_-<. >'=jf='< `. _ .' ,'--__--'. >> / . \ / \ /'-___-'\ / :| \ >> (_) . (_) / \ / \ (_) :| (_) >> \_-----'____--/ (_) (_) (_)_______(_) |___:|____| >> \___________/ |________| \_______/ |_________| >> _______________________________________________ >> 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 -- *Craig Lewis* Senior Systems Engineer Office +1.714.602.1309 Email clewis at centraldesktop.com <mailto:clewis at centraldesktop.com> *Central Desktop. Work together in ways you never thought possible.* Connect with us Website <http://www.centraldesktop.com/> | Twitter <http://www.twitter.com/centraldesktop> | Facebook <http://www.facebook.com/CentralDesktop> | LinkedIn <http://www.linkedin.com/groups?gid=147417> | Blog <http://cdblog.centraldesktop.com/> -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.ceph.com/pipermail/ceph-users-ceph.com/attachments/20140520/ad747ad2/attachment.htm>
 |