Re: [ceph-users] Add ssd's to hdd cluster, crush map class hdd update necessary?
This is actually not to nice, because this remapping is now causing a nearfull -Original Message- From: Dan van der Ster [mailto:d...@vanderster.com] Sent: woensdag 13 juni 2018 14:02 To: Marc Roos Cc: ceph-users Subject: Re: [ceph-users] Add ssd's to hdd cluster, crush map class hdd update necessary? See this thread: http://lists.ceph.com/pipermail/ceph-large-ceph.com/2018-April/000106.html http://lists.ceph.com/pipermail/ceph-large-ceph.com/2018-June/000113.html (Wido -- should we kill the ceph-large list??) On Wed, Jun 13, 2018 at 1:14 PM Marc Roos wrote: > > > I wonder if this is not a bug or so. Adding the class hdd, to an all > hdd cluster should not have such result that 60% of objects are moved > around. > > > pool fs_data.ec21 id 53 > 3866523/6247464 objects misplaced (61.889%) > recovery io 93089 kB/s, 22 objects/s > > > > > > -Original Message- > From: Marc Roos > Sent: woensdag 13 juni 2018 7:14 > To: ceph-users; k0ste > Subject: Re: [ceph-users] Add ssd's to hdd cluster, crush map class > hdd update necessary? > > I just added here 'class hdd' > > rule fs_data.ec21 { > id 4 > type erasure > min_size 3 > max_size 3 > step set_chooseleaf_tries 5 > step set_choose_tries 100 > step take default class hdd > step choose indep 0 type osd > step emit > } > > > -Original Message- > From: Konstantin Shalygin [mailto:k0...@k0ste.ru] > Sent: woensdag 13 juni 2018 12:30 > To: Marc Roos; ceph-users > Subject: *SPAM* Re: *SPAM* Re: [ceph-users] Add ssd's > to hdd cluster, crush map class hdd update necessary? > > On 06/13/2018 12:06 PM, Marc Roos wrote: > > Shit, I added this class and now everything start backfilling (10%) > > How is this possible, I only have hdd's? > > This is normal when you change your crush and placement rules. > Post your output, I will take a look > > ceph osd crush tree > ceph osd crush dump > ceph osd pool ls detail > > > > > > k > > > ___ > ceph-users mailing list > ceph-users@lists.ceph.com > http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com > > > ___ > ceph-users mailing list > ceph-users@lists.ceph.com > http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com ___ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
Re: [ceph-users] Add ssd's to hdd cluster, crush map class hdd update necessary?
Yes thanks I know, I will change it when I get extra an extra node. -Original Message- From: Paul Emmerich [mailto:paul.emmer...@croit.io] Sent: woensdag 13 juni 2018 16:33 To: Marc Roos Cc: ceph-users; k0ste Subject: Re: [ceph-users] Add ssd's to hdd cluster, crush map class hdd update necessary? 2018-06-13 7:13 GMT+02:00 Marc Roos : I just added here 'class hdd' rule fs_data.ec21 { id 4 type erasure min_size 3 max_size 3 step set_chooseleaf_tries 5 step set_choose_tries 100 step take default class hdd step choose indep 0 type osd step emit } somewhat off-topic, but: 2/1 erasure coding is usually a bad idea for the same reasons that size = 2 replicated pools are a bad idea. Paul -Original Message- From: Konstantin Shalygin [mailto:k0...@k0ste.ru] Sent: woensdag 13 juni 2018 12:30 To: Marc Roos; ceph-users Subject: *SPAM* Re: *SPAM* Re: [ceph-users] Add ssd's to hdd cluster, crush map class hdd update necessary? On 06/13/2018 12:06 PM, Marc Roos wrote: > Shit, I added this class and now everything start backfilling (10%) > How is this possible, I only have hdd's? This is normal when you change your crush and placement rules. Post your output, I will take a look ceph osd crush tree ceph osd crush dump ceph osd pool ls detail k ___ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com <http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com> -- Paul Emmerich Looking for help with your Ceph cluster? Contact us at https://croit.io croit GmbH Freseniusstr. 31h 81247 München www.croit.io Tel: +49 89 1896585 90 ___ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
Re: [ceph-users] Add ssd's to hdd cluster, crush map class hdd update necessary?
2018-06-13 7:13 GMT+02:00 Marc Roos : > I just added here 'class hdd' > > rule fs_data.ec21 { > id 4 > type erasure > min_size 3 > max_size 3 > step set_chooseleaf_tries 5 > step set_choose_tries 100 > step take default class hdd > step choose indep 0 type osd > step emit > } > somewhat off-topic, but: 2/1 erasure coding is usually a bad idea for the same reasons that size = 2 replicated pools are a bad idea. Paul > > > -Original Message- > From: Konstantin Shalygin [mailto:k0...@k0ste.ru] > Sent: woensdag 13 juni 2018 12:30 > To: Marc Roos; ceph-users > Subject: *****SPAM* Re: *****SPAM***** Re: [ceph-users] Add ssd's to > hdd cluster, crush map class hdd update necessary? > > On 06/13/2018 12:06 PM, Marc Roos wrote: > > Shit, I added this class and now everything start backfilling (10%) > > How is this possible, I only have hdd's? > > This is normal when you change your crush and placement rules. > Post your output, I will take a look > > ceph osd crush tree > ceph osd crush dump > ceph osd pool ls detail > > > > > > k > > > ___ > ceph-users mailing list > ceph-users@lists.ceph.com > http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com > -- Paul Emmerich Looking for help with your Ceph cluster? Contact us at https://croit.io croit GmbH Freseniusstr. 31h 81247 München www.croit.io Tel: +49 89 1896585 90 ___ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
Re: [ceph-users] Add ssd's to hdd cluster, crush map class hdd update necessary?
See this thread: http://lists.ceph.com/pipermail/ceph-large-ceph.com/2018-April/000106.html http://lists.ceph.com/pipermail/ceph-large-ceph.com/2018-June/000113.html (Wido -- should we kill the ceph-large list??) On Wed, Jun 13, 2018 at 1:14 PM Marc Roos wrote: > > > I wonder if this is not a bug or so. Adding the class hdd, to an all hdd > cluster should not have such result that 60% of objects are moved > around. > > > pool fs_data.ec21 id 53 > 3866523/6247464 objects misplaced (61.889%) > recovery io 93089 kB/s, 22 objects/s > > > > > > -Original Message- > From: Marc Roos > Sent: woensdag 13 juni 2018 7:14 > To: ceph-users; k0ste > Subject: Re: [ceph-users] Add ssd's to hdd cluster, crush map class hdd > update necessary? > > I just added here 'class hdd' > > rule fs_data.ec21 { > id 4 > type erasure > min_size 3 > max_size 3 > step set_chooseleaf_tries 5 > step set_choose_tries 100 > step take default class hdd > step choose indep 0 type osd > step emit > } > > > -Original Message- > From: Konstantin Shalygin [mailto:k0...@k0ste.ru] > Sent: woensdag 13 juni 2018 12:30 > To: Marc Roos; ceph-users > Subject: *SPAM* Re: *SPAM* Re: [ceph-users] Add ssd's to > hdd cluster, crush map class hdd update necessary? > > On 06/13/2018 12:06 PM, Marc Roos wrote: > > Shit, I added this class and now everything start backfilling (10%) > > How is this possible, I only have hdd's? > > This is normal when you change your crush and placement rules. > Post your output, I will take a look > > ceph osd crush tree > ceph osd crush dump > ceph osd pool ls detail > > > > > > k > > > ___ > ceph-users mailing list > ceph-users@lists.ceph.com > http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com > > > ___ > ceph-users mailing list > ceph-users@lists.ceph.com > http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com ___ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
Re: [ceph-users] Add ssd's to hdd cluster, crush map class hdd update necessary?
I wonder if this is not a bug or so. Adding the class hdd, to an all hdd cluster should not have such result that 60% of objects are moved around. pool fs_data.ec21 id 53 3866523/6247464 objects misplaced (61.889%) recovery io 93089 kB/s, 22 objects/s -Original Message- From: Marc Roos Sent: woensdag 13 juni 2018 7:14 To: ceph-users; k0ste Subject: Re: [ceph-users] Add ssd's to hdd cluster, crush map class hdd update necessary? I just added here 'class hdd' rule fs_data.ec21 { id 4 type erasure min_size 3 max_size 3 step set_chooseleaf_tries 5 step set_choose_tries 100 step take default class hdd step choose indep 0 type osd step emit } -Original Message- From: Konstantin Shalygin [mailto:k0...@k0ste.ru] Sent: woensdag 13 juni 2018 12:30 To: Marc Roos; ceph-users Subject: *SPAM* Re: *SPAM* Re: [ceph-users] Add ssd's to hdd cluster, crush map class hdd update necessary? On 06/13/2018 12:06 PM, Marc Roos wrote: > Shit, I added this class and now everything start backfilling (10%) > How is this possible, I only have hdd's? This is normal when you change your crush and placement rules. Post your output, I will take a look ceph osd crush tree ceph osd crush dump ceph osd pool ls detail k ___ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com ___ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
Re: [ceph-users] Add ssd's to hdd cluster, crush map class hdd update necessary?
I just added here 'class hdd' rule fs_data.ec21 { id 4 type erasure min_size 3 max_size 3 step set_chooseleaf_tries 5 step set_choose_tries 100 step take default class hdd step choose indep 0 type osd step emit } -Original Message- From: Konstantin Shalygin [mailto:k0...@k0ste.ru] Sent: woensdag 13 juni 2018 12:30 To: Marc Roos; ceph-users Subject: *SPAM* Re: *SPAM* Re: [ceph-users] Add ssd's to hdd cluster, crush map class hdd update necessary? On 06/13/2018 12:06 PM, Marc Roos wrote: > Shit, I added this class and now everything start backfilling (10%) > How is this possible, I only have hdd's? This is normal when you change your crush and placement rules. Post your output, I will take a look ceph osd crush tree ceph osd crush dump ceph osd pool ls detail k ___ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
Re: [ceph-users] Add ssd's to hdd cluster, crush map class hdd update necessary?
On 06/13/2018 09:01 AM, Marc Roos wrote: Yes but I already have some sort of test cluster with data in it. I don’t think there are commands to modify existing rules that are being used by pools. And the default replicated_ruleset doesn’t have a class specified. I also have an erasure code rule without any class definition for the file system. Yes, before migration from multi-root/classless crush to luminous+ classified crush you need to assign classified rulesets to your pools. This is safe to apply on production clusters, on EC pools too. k ___ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
Re: [ceph-users] Add ssd's to hdd cluster, crush map class hdd update necessary?
Yes but I already have some sort of test cluster with data in it. I don’t think there are commands to modify existing rules that are being used by pools. And the default replicated_ruleset doesn’t have a class specified. I also have an erasure code rule without any class definition for the file system. -Original Message- From: Konstantin Shalygin [mailto:k0...@k0ste.ru] Sent: woensdag 13 juni 2018 5:59 To: ceph-users@lists.ceph.com Cc: Marc Roos Subject: *SPAM* Re: [ceph-users] Add ssd's to hdd cluster, crush map class hdd update necessary? > Is it necessary to update the crush map with > > class hdd > > Before adding ssd's the cluster? Of course, if this osds of one root. It is not necessary to manually edit crush: ceph osd crush rule create-replicated replicated_hosts_hdd default host hdd ceph osd crush rule create-replicated replicated_hosts_nvme default host nvme k ___ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
Re: [ceph-users] Add ssd's to hdd cluster, crush map class hdd update necessary?
Is it necessary to update the crush map with class hdd Before adding ssd's the cluster? Of course, if this osds of one root. It is not necessary to manually edit crush: ceph osd crush rule create-replicated replicated_hosts_hdd default host hdd ceph osd crush rule create-replicated replicated_hosts_nvme default host nvme k ___ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
Re: [ceph-users] Add ssd's to hdd cluster, crush map class hdd update necessary?
Eg # rules rule replicated_ruleset { id 0 type replicated min_size 1 max_size 10 step take default step chooseleaf firstn 0 type host step emit } To # rules rule replicated_ruleset { id 0 type replicated min_size 1 max_size 10 step take default class hdd step chooseleaf firstn 0 type host step emit } And rule fs_data.ec21 { id 4 type erasure min_size 3 max_size 3 step set_chooseleaf_tries 5 step set_choose_tries 100 step take default step choose indep 0 type osd step emit } To rule fs_data.ec21 { id 4 type erasure min_size 3 max_size 3 step set_chooseleaf_tries 5 step set_choose_tries 100 step take default class hdd step choose indep 0 type osd step emit } -Original Message- From: Marc Roos Sent: dinsdag 12 juni 2018 17:07 To: ceph-users Subject: [ceph-users] Add ssd's to hdd cluster, crush map class hdd update necessary? Is it necessary to update the crush map with class hdd Before adding ssd's the cluster? ___ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com ___ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
[ceph-users] Add ssd's to hdd cluster, crush map class hdd update necessary?
Is it necessary to update the crush map with class hdd Before adding ssd's the cluster? ___ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com