[ceph-users] PG Balancer Upmap mode not working
Hi,the docs say the upmap mode is trying to achieve perfect distribution as to have equal amount of PGs/OSD.This is what I got(v14.2.4): 0 ssd 3.49219 1.0 3.5 TiB 794 GiB 753 GiB 38 GiB 3.4 GiB 2.7 TiB 22.20 0.32 82 up 1 ssd 3.49219 1.0 3.5 TiB 800 GiB 751 GiB 45 GiB 3.7 GiB 2.7 TiB 22.37 0.33 84 up 2 ssd 3.49219 1.0 3.5 TiB 846 GiB 792 GiB 50 GiB 3.6 GiB 2.7 TiB 23.66 0.35 88 up 3 ssd 3.49219 1.0 3.5 TiB 812 GiB 776 GiB 33 GiB 3.3 GiB 2.7 TiB 22.71 0.33 85 up 4 ssd 3.49219 1.0 3.5 TiB 768 GiB 730 GiB 34 GiB 4.1 GiB 2.7 TiB 21.47 0.31 83 up 6 ssd 3.49219 1.0 3.5 TiB 765 GiB 731 GiB 31 GiB 3.3 GiB 2.7 TiB 21.40 0.31 82 up 8 ssd 3.49219 1.0 3.5 TiB 872 GiB 828 GiB 41 GiB 3.2 GiB 2.6 TiB 24.40 0.36 85 up 10 ssd 3.49219 1.0 3.5 TiB 789 GiB 743 GiB 42 GiB 3.3 GiB 2.7 TiB 22.05 0.32 82 up 5 ssd 3.49219 1.0 3.5 TiB 719 GiB 683 GiB 32 GiB 3.9 GiB 2.8 TiB 20.12 0.29 78 up 7 ssd 3.49219 1.0 3.5 TiB 741 GiB 698 GiB 39 GiB 3.8 GiB 2.8 TiB 20.73 0.30 79 up 9 ssd 3.49219 1.0 3.5 TiB 709 GiB 664 GiB 41 GiB 3.5 GiB 2.8 TiB 19.82 0.29 78 up 11 ssd 3.49219 1.0 3.5 TiB 858 GiB 834 GiB 22 GiB 2.4 GiB 2.7 TiB 23.99 0.35 82 up 101 ssd 3.49219 1.0 3.5 TiB 815 GiB 774 GiB 38 GiB 3.5 GiB 2.7 TiB 22.80 0.33 80 up 103 ssd 3.49219 1.0 3.5 TiB 827 GiB 783 GiB 40 GiB 3.3 GiB 2.7 TiB 23.11 0.34 81 up 105 ssd 3.49219 1.0 3.5 TiB 797 GiB 759 GiB 36 GiB 2.5 GiB 2.7 TiB 22.30 0.33 81 up 107 ssd 3.49219 1.0 3.5 TiB 840 GiB 788 GiB 50 GiB 2.8 GiB 2.7 TiB 23.50 0.34 83 up 100 ssd 3.49219 1.0 3.5 TiB 728 GiB 678 GiB 47 GiB 2.4 GiB 2.8 TiB 20.36 0.30 78 up 102 ssd 3.49219 1.0 3.5 TiB 764 GiB 750 GiB 12 GiB 2.2 GiB 2.7 TiB 21.37 0.31 76 up 104 ssd 3.49219 1.0 3.5 TiB 795 GiB 761 GiB 31 GiB 2.5 GiB 2.7 TiB 22.22 0.33 78 up 106 ssd 3.49219 1.0 3.5 TiB 730 GiB 665 GiB 62 GiB 2.8 GiB 2.8 TiB 20.41 0.30 78 up 108 ssd 3.49219 1.0 3.5 TiB 849 GiB 808 GiB 38 GiB 2.5 GiB 2.7 TiB 23.73 0.35 92 up 109 ssd 3.49219 1.0 3.5 TiB 798 GiB 754 GiB 41 GiB 2.7 GiB 2.7 TiB 22.30 0.33 83 up 110 ssd 3.49219 1.0 3.5 TiB 840 GiB 810 GiB 28 GiB 2.4 GiB 2.7 TiB 23.49 0.34 85 up 111 ssd 3.49219 1.0 3.5 TiB 788 GiB 741 GiB 45 GiB 2.5 GiB 2.7 TiB 22.04 0.32 85 up PG's are badly distributed.ceph balancer status { "active": true, "plans": [], "mode": "upmap" } It is because of this? health: HEALTH_WARN Failed to send data to Zabbix 1 subtrees have overcommitted pool target_size_bytes 1 subtrees have overcommitted pool target_size_ratio Any ideas why its not working? ___ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
Re: [ceph-users] PG Balancer Upmap mode not working
On 12/7/19 11:42 AM, Philippe D'Anjou wrote: > Hi, > the docs say the upmap mode is trying to achieve perfect distribution as > to have equal amount of PGs/OSD. > This is what I got(v14.2.4): > > 0 ssd 3.49219 1.0 3.5 TiB 794 GiB 753 GiB 38 GiB 3.4 GiB 2.7 > TiB 22.20 0.32 82 up > 1 ssd 3.49219 1.0 3.5 TiB 800 GiB 751 GiB 45 GiB 3.7 GiB 2.7 > TiB 22.37 0.33 84 up > 2 ssd 3.49219 1.0 3.5 TiB 846 GiB 792 GiB 50 GiB 3.6 GiB 2.7 > TiB 23.66 0.35 88 up > 3 ssd 3.49219 1.0 3.5 TiB 812 GiB 776 GiB 33 GiB 3.3 GiB 2.7 > TiB 22.71 0.33 85 up > 4 ssd 3.49219 1.0 3.5 TiB 768 GiB 730 GiB 34 GiB 4.1 GiB 2.7 > TiB 21.47 0.31 83 up > 6 ssd 3.49219 1.0 3.5 TiB 765 GiB 731 GiB 31 GiB 3.3 GiB 2.7 > TiB 21.40 0.31 82 up > 8 ssd 3.49219 1.0 3.5 TiB 872 GiB 828 GiB 41 GiB 3.2 GiB 2.6 > TiB 24.40 0.36 85 up > 10 ssd 3.49219 1.0 3.5 TiB 789 GiB 743 GiB 42 GiB 3.3 GiB 2.7 > TiB 22.05 0.32 82 up > 5 ssd 3.49219 1.0 3.5 TiB 719 GiB 683 GiB 32 GiB 3.9 GiB 2.8 > TiB 20.12 0.29 78 up > 7 ssd 3.49219 1.0 3.5 TiB 741 GiB 698 GiB 39 GiB 3.8 GiB 2.8 > TiB 20.73 0.30 79 up > 9 ssd 3.49219 1.0 3.5 TiB 709 GiB 664 GiB 41 GiB 3.5 GiB 2.8 > TiB 19.82 0.29 78 up > 11 ssd 3.49219 1.0 3.5 TiB 858 GiB 834 GiB 22 GiB 2.4 GiB 2.7 > TiB 23.99 0.35 82 up > 101 ssd 3.49219 1.0 3.5 TiB 815 GiB 774 GiB 38 GiB 3.5 GiB 2.7 > TiB 22.80 0.33 80 up > 103 ssd 3.49219 1.0 3.5 TiB 827 GiB 783 GiB 40 GiB 3.3 GiB 2.7 > TiB 23.11 0.34 81 up > 105 ssd 3.49219 1.0 3.5 TiB 797 GiB 759 GiB 36 GiB 2.5 GiB 2.7 > TiB 22.30 0.33 81 up > 107 ssd 3.49219 1.0 3.5 TiB 840 GiB 788 GiB 50 GiB 2.8 GiB 2.7 > TiB 23.50 0.34 83 up > 100 ssd 3.49219 1.0 3.5 TiB 728 GiB 678 GiB 47 GiB 2.4 GiB 2.8 > TiB 20.36 0.30 78 up > 102 ssd 3.49219 1.0 3.5 TiB 764 GiB 750 GiB 12 GiB 2.2 GiB 2.7 > TiB 21.37 0.31 76 up > 104 ssd 3.49219 1.0 3.5 TiB 795 GiB 761 GiB 31 GiB 2.5 GiB 2.7 > TiB 22.22 0.33 78 up > 106 ssd 3.49219 1.0 3.5 TiB 730 GiB 665 GiB 62 GiB 2.8 GiB 2.8 > TiB 20.41 0.30 78 up > 108 ssd 3.49219 1.0 3.5 TiB 849 GiB 808 GiB 38 GiB 2.5 GiB 2.7 > TiB 23.73 0.35 92 up > 109 ssd 3.49219 1.0 3.5 TiB 798 GiB 754 GiB 41 GiB 2.7 GiB 2.7 > TiB 22.30 0.33 83 up > 110 ssd 3.49219 1.0 3.5 TiB 840 GiB 810 GiB 28 GiB 2.4 GiB 2.7 > TiB 23.49 0.34 85 up > 111 ssd 3.49219 1.0 3.5 TiB 788 GiB 741 GiB 45 GiB 2.5 GiB 2.7 > TiB 22.04 0.32 85 up > > PG's are badly distributed. From what information do you draw that conclusion? You use about 22% on all OSDs. I suggest that you increase your PGs to at least 100 per OSD, that will make distribution even better. Wido > ceph balancer status > { > "active": true, > "plans": [], > "mode": "upmap" > } > > It is because of this? > health: HEALTH_WARN > Failed to send data to Zabbix > 1 subtrees have overcommitted pool target_size_bytes > 1 subtrees have overcommitted pool target_size_ratio > > > Any ideas why its not working? > > > ___ > 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] PG Balancer Upmap mode not working
@Wido Den Hollander That doesn't explain why its between 76 and 92 PGs, that's major not equal. Raising PGs to 100 is an old statement anyway, anything 60+ should be fine. Not an excuse for distribution failure in this case.I am expecting more or less equal PGs/OSD ___ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
Re: [ceph-users] PG Balancer Upmap mode not working
On 12/7/19 1:42 PM, Philippe D'Anjou wrote: > @Wido Den Hollander > > That doesn't explain why its between 76 and 92 PGs, that's major not equal. The balancer will balance the PGs so that all OSDs have an almost equal data usage. It doesn't balance that all OSDs have an equal amount of PGs. The end goal is to make sure all OSDs are filled equally. > Raising PGs to 100 is an old statement anyway, anything 60+ should be > fine. Not an excuse for distribution failure in this case. > I am expecting more or less equal PGs/OSD That will not happen. Objects are distributed over PGs based on the hash of their name and not their size. The distribution of objects over PGs is almost perfect. However, objects will vary in size. If you are low on the amount of PGs you'll have PGs which are bigger then others. This causes data distribution to be off. Therefor it can help to increase the amount of PGs as that will result in a better distribution. PGs will be less varied in size and makes them easier to balance. In addition you gain more parallel performance and suffer less from PG contention. As disks (or SSDs) become bigger and bigger you can benefit by having more PGs. Hope this explains it :-) Wido ___ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
[ceph-users] PG Balancer Upmap mode not working
@Wido Den Hollander First of all the docs say: " In most cases, this distribution is “perfect,” whichan equal number of PGs on each OSD (+/-1 PG, since they might notdivide evenly)."Either this is just false information or very badly stated. I increased PGs and see no difference. I pointed out MULTIPLE times that Nautilus has major flaws in the data distribution but nobody seems to listen to me. Not sure how much more evidence I have to show. 0 ssd 3.49219 1.0 3.5 TiB 715 GiB 674 GiB 37 GiB 3.9 GiB 2.8 TiB 19.99 0.29 147 up 1 ssd 3.49219 1.0 3.5 TiB 724 GiB 672 GiB 49 GiB 3.8 GiB 2.8 TiB 20.25 0.30 146 up 2 ssd 3.49219 1.0 3.5 TiB 736 GiB 681 GiB 50 GiB 4.4 GiB 2.8 TiB 20.57 0.30 150 up 3 ssd 3.49219 1.0 3.5 TiB 712 GiB 676 GiB 33 GiB 3.5 GiB 2.8 TiB 19.92 0.29 146 up 4 ssd 3.49219 1.0 3.5 TiB 752 GiB 714 GiB 34 GiB 4.6 GiB 2.8 TiB 21.03 0.31 156 up 6 ssd 3.49219 1.0 3.5 TiB 710 GiB 671 GiB 35 GiB 3.8 GiB 2.8 TiB 19.85 0.29 146 up 8 ssd 3.49219 1.0 3.5 TiB 781 GiB 738 GiB 40 GiB 3.7 GiB 2.7 TiB 21.85 0.32 156 up 10 ssd 3.49219 1.0 3.5 TiB 728 GiB 682 GiB 42 GiB 4.0 GiB 2.8 TiB 20.35 0.30 146 up 5 ssd 3.49219 1.0 3.5 TiB 664 GiB 628 GiB 32 GiB 4.3 GiB 2.8 TiB 18.58 0.27 141 up 7 ssd 3.49219 1.0 3.5 TiB 656 GiB 613 GiB 39 GiB 4.0 GiB 2.9 TiB 18.35 0.27 136 up 9 ssd 3.49219 1.0 3.5 TiB 632 GiB 586 GiB 41 GiB 4.4 GiB 2.9 TiB 17.67 0.26 131 up 11 ssd 3.49219 1.0 3.5 TiB 725 GiB 701 GiB 22 GiB 2.6 GiB 2.8 TiB 20.28 0.30 138 up 101 ssd 3.49219 1.0 3.5 TiB 755 GiB 713 GiB 38 GiB 3.9 GiB 2.8 TiB 21.11 0.31 146 up 103 ssd 3.49219 1.0 3.5 TiB 761 GiB 718 GiB 40 GiB 3.6 GiB 2.7 TiB 21.29 0.31 150 up 105 ssd 3.49219 1.0 3.5 TiB 715 GiB 676 GiB 36 GiB 2.6 GiB 2.8 TiB 19.99 0.29 148 up 107 ssd 3.49219 1.0 3.5 TiB 760 GiB 706 GiB 50 GiB 3.2 GiB 2.8 TiB 21.24 0.31 147 up 100 ssd 3.49219 1.0 3.5 TiB 724 GiB 674 GiB 47 GiB 2.5 GiB 2.8 TiB 20.25 0.30 144 up 102 ssd 3.49219 1.0 3.5 TiB 669 GiB 654 GiB 12 GiB 2.3 GiB 2.8 TiB 18.71 0.27 141 up 104 ssd 3.49219 1.0 3.5 TiB 721 GiB 687 GiB 31 GiB 3.0 GiB 2.8 TiB 20.16 0.30 144 up 106 ssd 3.49219 1.0 3.5 TiB 715 GiB 646 GiB 65 GiB 3.8 GiB 2.8 TiB 19.99 0.29 143 up 108 ssd 3.49219 1.0 3.5 TiB 729 GiB 691 GiB 36 GiB 2.6 GiB 2.8 TiB 20.38 0.30 156 up 109 ssd 3.49219 1.0 3.5 TiB 732 GiB 684 GiB 45 GiB 3.0 GiB 2.8 TiB 20.47 0.30 146 up 110 ssd 3.49219 1.0 3.5 TiB 773 GiB 743 GiB 28 GiB 2.7 GiB 2.7 TiB 21.63 0.32 154 up 111 ssd 3.49219 1.0 3.5 TiB 708 GiB 660 GiB 45 GiB 2.7 GiB 2.8 TiB 19.78 0.29 146 up The % fillrate is no different than before, fluctuates hard. ___ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
Re: [ceph-users] PG Balancer Upmap mode not working
On 12/7/19 3:39 PM, Philippe D'Anjou wrote: > @Wido Den Hollander > > First of all the docs say: "In most cases, this distribution is > “perfect,” which an equal number of PGs on each OSD (+/-1 PG, since they > might not divide evenly)." > Either this is just false information or very badly stated. Might be both. But what are you trying to achieve? PGs will never be equally sized because objects vary in size. The end result, I assume, is that you have equally filled OSDs. > > I increased PGs and see no difference. > > I pointed out MULTIPLE times that Nautilus has major flaws in the data > distribution but nobody seems to listen to me. Not sure how much more > evidence I have to show. > What has changed? Because this can only change if Nautilus had a CRUSH algorithm change which it didn't. Upgrading from Mimic nor Luminous causes a major shift in data. > 0 ssd 3.49219 1.0 3.5 TiB 715 GiB 674 GiB 37 GiB 3.9 GiB 2.8 > TiB 19.99 0.29 147 up > 1 ssd 3.49219 1.0 3.5 TiB 724 GiB 672 GiB 49 GiB 3.8 GiB 2.8 > TiB 20.25 0.30 146 up > 2 ssd 3.49219 1.0 3.5 TiB 736 GiB 681 GiB 50 GiB 4.4 GiB 2.8 > TiB 20.57 0.30 150 up > 3 ssd 3.49219 1.0 3.5 TiB 712 GiB 676 GiB 33 GiB 3.5 GiB 2.8 > TiB 19.92 0.29 146 up > 4 ssd 3.49219 1.0 3.5 TiB 752 GiB 714 GiB 34 GiB 4.6 GiB 2.8 > TiB 21.03 0.31 156 up > 6 ssd 3.49219 1.0 3.5 TiB 710 GiB 671 GiB 35 GiB 3.8 GiB 2.8 > TiB 19.85 0.29 146 up > 8 ssd 3.49219 1.0 3.5 TiB 781 GiB 738 GiB 40 GiB 3.7 GiB 2.7 > TiB 21.85 0.32 156 up > 10 ssd 3.49219 1.0 3.5 TiB 728 GiB 682 GiB 42 GiB 4.0 GiB 2.8 > TiB 20.35 0.30 146 up > 5 ssd 3.49219 1.0 3.5 TiB 664 GiB 628 GiB 32 GiB 4.3 GiB 2.8 > TiB 18.58 0.27 141 up > 7 ssd 3.49219 1.0 3.5 TiB 656 GiB 613 GiB 39 GiB 4.0 GiB 2.9 > TiB 18.35 0.27 136 up > 9 ssd 3.49219 1.0 3.5 TiB 632 GiB 586 GiB 41 GiB 4.4 GiB 2.9 > TiB 17.67 0.26 131 up > 11 ssd 3.49219 1.0 3.5 TiB 725 GiB 701 GiB 22 GiB 2.6 GiB 2.8 > TiB 20.28 0.30 138 up > 101 ssd 3.49219 1.0 3.5 TiB 755 GiB 713 GiB 38 GiB 3.9 GiB 2.8 > TiB 21.11 0.31 146 up > 103 ssd 3.49219 1.0 3.5 TiB 761 GiB 718 GiB 40 GiB 3.6 GiB 2.7 > TiB 21.29 0.31 150 up > 105 ssd 3.49219 1.0 3.5 TiB 715 GiB 676 GiB 36 GiB 2.6 GiB 2.8 > TiB 19.99 0.29 148 up > 107 ssd 3.49219 1.0 3.5 TiB 760 GiB 706 GiB 50 GiB 3.2 GiB 2.8 > TiB 21.24 0.31 147 up > 100 ssd 3.49219 1.0 3.5 TiB 724 GiB 674 GiB 47 GiB 2.5 GiB 2.8 > TiB 20.25 0.30 144 up > 102 ssd 3.49219 1.0 3.5 TiB 669 GiB 654 GiB 12 GiB 2.3 GiB 2.8 > TiB 18.71 0.27 141 up > 104 ssd 3.49219 1.0 3.5 TiB 721 GiB 687 GiB 31 GiB 3.0 GiB 2.8 > TiB 20.16 0.30 144 up > 106 ssd 3.49219 1.0 3.5 TiB 715 GiB 646 GiB 65 GiB 3.8 GiB 2.8 > TiB 19.99 0.29 143 up > 108 ssd 3.49219 1.0 3.5 TiB 729 GiB 691 GiB 36 GiB 2.6 GiB 2.8 > TiB 20.38 0.30 156 up > 109 ssd 3.49219 1.0 3.5 TiB 732 GiB 684 GiB 45 GiB 3.0 GiB 2.8 > TiB 20.47 0.30 146 up > 110 ssd 3.49219 1.0 3.5 TiB 773 GiB 743 GiB 28 GiB 2.7 GiB 2.7 > TiB 21.63 0.32 154 up > 111 ssd 3.49219 1.0 3.5 TiB 708 GiB 660 GiB 45 GiB 2.7 GiB 2.8 > TiB 19.78 0.29 146 up > > The % fillrate is no different than before, fluctuates hard. All OSDs are very close to 20%, that's very good. What is the real problem here? Wido ___ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
[ceph-users] PG Balancer Upmap mode not working
I never had those issues with Luminous, never once, since Nautilus this is a constant headache.My issue is that I have OSDs that are over 85% whilst others are at 63%. My issue is that every time I do a rebalance or add new disks ceph moves PGs on near full OSDs and almost causes pool failures. My STDDEV: 21.31 ...it's a joke. It's simply not acceptable to deal with nearfull OSDs whilst others are half empty. ___ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
[ceph-users] Cephfs metadata fix tool
Our Jewel cluster is exhibiting some similar issues to the one in this thread [0] and it was indicated that a tool would need to be written to fix that kind of corruption. Has the tool been written? How would I go about repair this 16EB directories that won't delete? Thank you, Robert LeBlanc [0] https://www.spinics.net/lists/ceph-users/msg31598.html Robert LeBlanc PGP Fingerprint 79A2 9CA4 6CC4 45DD A904 C70E E654 3BB2 FA62 B9F1 ___ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com