Re: should I increase the amount of PGs?

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Thanks.

Decreasing the max deviation to 2 or 1 should help in your case. This
option controls when the balancer stops trying to move PGs around -- by
default it stops when the deviation from the mean is 5. Yes this is too
large IMO -- all of our clusters have this set to 1.

And given that you have some OSDs with more than 200 PGs, you definitely
shouldn't increase the num PGs.

But anyway with your mixed device sizes it might be challenging to make a
perfectly uniform distribution. Give it a try with 1 though, and let us
know how it goes.

.. Dan





On Sat, Mar 13, 2021, 4:11 PM Boris Behrens <bb@xxxxxxxxx> wrote:

> Hi Dan,
>
> upmap_max_deviation is default (5) in our cluster. Is 1 the recommended
> deviation?
>
> I added the whole ceph osd df tree, (I need to remove some OSDs and readd
> them as bluestore with SSD, so 69, 73 and 82 are a bit off now. I also
> reweighted to try to get the %USE mitigated).
>
> I will increase the mgr debugging to see what is the problem.
>
> [root@s3db1 ~]# ceph osd df tree
> ID  CLASS WEIGHT    REWEIGHT SIZE    RAW USE DATA    OMAP    META    AVAIL
>   %USE  VAR  PGS STATUS TYPE NAME
>  -1       673.54224        - 659 TiB 491 TiB 464 TiB  96 GiB 1.2 TiB 168
> TiB 74.57 1.00   -        root default
>  -2        58.30331        -  44 TiB  22 TiB  17 TiB 5.7 GiB  38 GiB  22
> TiB 49.82 0.67   -            host s3db1
>  23   hdd  14.65039  1.00000  15 TiB 1.8 TiB 1.7 TiB 156 MiB 4.4 GiB  13
> TiB 12.50 0.17 101     up         osd.23
>  69   hdd  14.55269        0     0 B     0 B     0 B     0 B     0 B     0
> B     0    0  11     up         osd.69
>  73   hdd  14.55269  1.00000  15 TiB  10 TiB  10 TiB 6.1 MiB  33 GiB 4.2
> TiB 71.15 0.95 107     up         osd.73
>  79   hdd   3.63689  1.00000 3.6 TiB 2.9 TiB 747 GiB 2.0 GiB     0 B 747
> GiB 79.94 1.07  52     up         osd.79
>  80   hdd   3.63689  1.00000 3.6 TiB 2.6 TiB 1.0 TiB 1.9 GiB     0 B 1.0
> TiB 71.61 0.96  58     up         osd.80
>  81   hdd   3.63689  1.00000 3.6 TiB 2.2 TiB 1.5 TiB 1.1 GiB     0 B 1.5
> TiB 60.07 0.81  55     up         osd.81
>  82   hdd   3.63689  1.00000 3.6 TiB 1.9 TiB 1.7 TiB 536 MiB     0 B 1.7
> TiB 52.68 0.71  30     up         osd.82
> -11        50.94173        -  51 TiB  38 TiB  38 TiB 3.7 GiB 100 GiB  13
> TiB 74.69 1.00   -            host s3db10
>  63   hdd   7.27739  1.00000 7.3 TiB 5.5 TiB 5.5 TiB 616 MiB  14 GiB 1.7
> TiB 76.04 1.02  92     up         osd.63
>  64   hdd   7.27739  1.00000 7.3 TiB 5.5 TiB 5.5 TiB 820 MiB  15 GiB 1.8
> TiB 75.54 1.01 101     up         osd.64
>  65   hdd   7.27739  1.00000 7.3 TiB 5.3 TiB 5.3 TiB 109 MiB  14 GiB 2.0
> TiB 73.17 0.98 105     up         osd.65
>  66   hdd   7.27739  1.00000 7.3 TiB 5.8 TiB 5.8 TiB 423 MiB  15 GiB 1.4
> TiB 80.38 1.08  98     up         osd.66
>  67   hdd   7.27739  1.00000 7.3 TiB 5.1 TiB 5.1 TiB 572 MiB  14 GiB 2.2
> TiB 70.10 0.94 100     up         osd.67
>  68   hdd   7.27739  1.00000 7.3 TiB 5.3 TiB 5.3 TiB 630 MiB  13 GiB 2.0
> TiB 72.88 0.98 107     up         osd.68
>  70   hdd   7.27739  1.00000 7.3 TiB 5.4 TiB 5.4 TiB 648 MiB  14 GiB 1.8
> TiB 74.73 1.00 102     up         osd.70
> -12        50.99052        -  51 TiB  39 TiB  39 TiB 2.9 GiB  99 GiB  12
> TiB 77.24 1.04   -            host s3db11
>  46   hdd   7.27739  1.00000 7.3 TiB 5.7 TiB 5.7 TiB 102 MiB  15 GiB 1.5
> TiB 78.91 1.06  97     up         osd.46
>  47   hdd   7.27739  1.00000 7.3 TiB 5.2 TiB 5.2 TiB  61 MiB  13 GiB 2.1
> TiB 71.47 0.96  96     up         osd.47
>  48   hdd   7.27739  1.00000 7.3 TiB 6.1 TiB 6.1 TiB 853 MiB  15 GiB 1.2
> TiB 83.46 1.12 109     up         osd.48
>  49   hdd   7.27739  1.00000 7.3 TiB 5.7 TiB 5.7 TiB 708 MiB  15 GiB 1.5
> TiB 78.96 1.06  98     up         osd.49
>  50   hdd   7.27739  1.00000 7.3 TiB 5.9 TiB 5.8 TiB 472 MiB  15 GiB 1.4
> TiB 80.40 1.08 102     up         osd.50
>  51   hdd   7.27739  1.00000 7.3 TiB 5.9 TiB 5.9 TiB 729 MiB  15 GiB 1.3
> TiB 81.70 1.10 110     up         osd.51
>  72   hdd   7.32619  1.00000 7.3 TiB 4.8 TiB 4.8 TiB  91 MiB  12 GiB 2.5
> TiB 65.82 0.88  89     up         osd.72
> -37        58.55478        -  59 TiB  46 TiB  46 TiB 5.0 GiB 124 GiB  12
> TiB 79.04 1.06   -            host s3db12
>  19   hdd   3.68750  1.00000 3.7 TiB 3.1 TiB 3.1 TiB 462 MiB 8.2 GiB 559
> GiB 85.18 1.14  55     up         osd.19
>  71   hdd   3.68750  1.00000 3.7 TiB 2.9 TiB 2.8 TiB 3.9 MiB 7.8 GiB 825
> GiB 78.14 1.05  50     up         osd.71
>  75   hdd   3.68750  1.00000 3.7 TiB 3.1 TiB 3.1 TiB 576 MiB 8.3 GiB 555
> GiB 85.29 1.14  57     up         osd.75
>  76   hdd   3.68750  1.00000 3.7 TiB 3.2 TiB 3.1 TiB 239 MiB 9.3 GiB 501
> GiB 86.73 1.16  50     up         osd.76
>  77   hdd  14.60159  1.00000  15 TiB  11 TiB  11 TiB 880 MiB  30 GiB 3.6
> TiB 75.57 1.01 202     up         osd.77
>  78   hdd  14.60159  1.00000  15 TiB  11 TiB  11 TiB 1.0 GiB  30 GiB 3.4
> TiB 76.65 1.03 196     up         osd.78
>  83   hdd  14.60159  1.00000  15 TiB  12 TiB  12 TiB 1.8 GiB  31 GiB 2.9
> TiB 80.04 1.07 223     up         osd.83
>  -3        58.49872        -  58 TiB  43 TiB  38 TiB 8.1 GiB  91 GiB  16
> TiB 73.15 0.98   -            host s3db2
>   1   hdd  14.65039  1.00000  15 TiB  11 TiB  11 TiB 3.1 GiB  38 GiB 3.6
> TiB 75.52 1.01 194     up         osd.1
>   3   hdd   3.63689  1.00000 3.6 TiB 2.2 TiB 1.4 TiB 418 MiB     0 B 1.4
> TiB 60.94 0.82  52     up         osd.3
>   4   hdd   3.63689  0.89999 3.6 TiB 3.2 TiB 401 GiB 845 MiB     0 B 401
> GiB 89.23 1.20  53     up         osd.4
>   5   hdd   3.63689  1.00000 3.6 TiB 2.3 TiB 1.3 TiB 437 MiB     0 B 1.3
> TiB 62.88 0.84  51     up         osd.5
>   6   hdd   3.63689  1.00000 3.6 TiB 2.0 TiB 1.7 TiB 1.8 GiB     0 B 1.7
> TiB 54.51 0.73  47     up         osd.6
>   7   hdd  14.65039  1.00000  15 TiB  11 TiB  11 TiB 493 MiB  26 GiB 3.8
> TiB 73.90 0.99 185     up         osd.7
>  74   hdd  14.65039  1.00000  15 TiB  11 TiB  11 TiB 1.1 GiB  27 GiB 3.5
> TiB 76.27 1.02 208     up         osd.74
>  -4        58.49872        -  58 TiB  43 TiB  37 TiB  33 GiB  86 GiB  15
> TiB 74.05 0.99   -            host s3db3
>   2   hdd  14.65039  1.00000  15 TiB  11 TiB  11 TiB 850 MiB  26 GiB 4.0
> TiB 72.78 0.98 203     up         osd.2
>   9   hdd  14.65039  1.00000  15 TiB  11 TiB  11 TiB 8.3 GiB  33 GiB 3.6
> TiB 75.62 1.01 189     up         osd.9
>  10   hdd  14.65039  1.00000  15 TiB  11 TiB  11 TiB 663 MiB  28 GiB 3.5
> TiB 76.34 1.02 211     up         osd.10
>  12   hdd   3.63689  1.00000 3.6 TiB 2.4 TiB 1.2 TiB 633 MiB     0 B 1.2
> TiB 66.22 0.89  44     up         osd.12
>  13   hdd   3.63689  1.00000 3.6 TiB 2.9 TiB 720 GiB 2.3 GiB     0 B 720
> GiB 80.66 1.08  66     up         osd.13
>  14   hdd   3.63689  1.00000 3.6 TiB 3.1 TiB 552 GiB  18 GiB     0 B 552
> GiB 85.18 1.14  60     up         osd.14
>  15   hdd   3.63689  1.00000 3.6 TiB 2.0 TiB 1.7 TiB 2.1 GiB     0 B 1.7
> TiB 53.72 0.72  44     up         osd.15
>  -5        58.49872        -  58 TiB  45 TiB  37 TiB 7.2 GiB  99 GiB  14
> TiB 76.37 1.02   -            host s3db4
>  11   hdd  14.65039  1.00000  15 TiB  12 TiB  12 TiB 897 MiB  28 GiB 2.8
> TiB 81.15 1.09 205     up         osd.11
>  17   hdd  14.65039  1.00000  15 TiB  11 TiB  11 TiB 1.2 GiB  27 GiB 3.6
> TiB 75.38 1.01 211     up         osd.17
>  18   hdd  14.65039  1.00000  15 TiB  11 TiB  11 TiB 965 MiB  44 GiB 4.0
> TiB 72.86 0.98 188     up         osd.18
>  20   hdd   3.63689  1.00000 3.6 TiB 2.9 TiB 796 GiB 529 MiB     0 B 796
> GiB 78.63 1.05  66     up         osd.20
>  21   hdd   3.63689  1.00000 3.6 TiB 2.6 TiB 1.1 TiB 2.1 GiB     0 B 1.1
> TiB 70.32 0.94  47     up         osd.21
>  22   hdd   3.63689  1.00000 3.6 TiB 2.9 TiB 802 GiB 882 MiB     0 B 802
> GiB 78.47 1.05  58     up         osd.22
>  24   hdd   3.63689  1.00000 3.6 TiB 2.8 TiB 856 GiB 645 MiB     0 B 856
> GiB 77.01 1.03  47     up         osd.24
>  -6        58.89636        -  59 TiB  44 TiB  44 TiB 2.4 GiB 111 GiB  15
> TiB 75.22 1.01   -            host s3db5
>   0   hdd   3.73630  1.00000 3.7 TiB 2.4 TiB 2.3 TiB  70 MiB 6.6 GiB 1.3
> TiB 65.00 0.87  48     up         osd.0
>  25   hdd   3.73630  1.00000 3.7 TiB 2.4 TiB 2.3 TiB 5.3 MiB 6.6 GiB 1.4
> TiB 63.86 0.86  41     up         osd.25
>  26   hdd   3.73630  1.00000 3.7 TiB 2.9 TiB 2.8 TiB 181 MiB 7.6 GiB 862
> GiB 77.47 1.04  48     up         osd.26
>  27   hdd   3.73630  1.00000 3.7 TiB 2.3 TiB 2.2 TiB 7.0 MiB 6.1 GiB 1.5
> TiB 61.00 0.82  48     up         osd.27
>  28   hdd  14.65039  1.00000  15 TiB  12 TiB  12 TiB 937 MiB  30 GiB 2.8
> TiB 81.19 1.09 203     up         osd.28
>  29   hdd  14.65039  1.00000  15 TiB  11 TiB  11 TiB 536 MiB  26 GiB 3.8
> TiB 73.95 0.99 200     up         osd.29
>  30   hdd  14.65039  1.00000  15 TiB  12 TiB  11 TiB 744 MiB  28 GiB 3.1
> TiB 79.07 1.06 207     up         osd.30
>  -7        58.89636        -  59 TiB  44 TiB  44 TiB  14 GiB 122 GiB  14
> TiB 75.41 1.01   -            host s3db6
>  32   hdd   3.73630  1.00000 3.7 TiB 3.1 TiB 3.0 TiB  16 MiB 8.2 GiB 622
> GiB 83.74 1.12  65     up         osd.32
>  33   hdd   3.73630  0.79999 3.7 TiB 3.0 TiB 2.9 TiB  14 MiB 8.1 GiB 740
> GiB 80.67 1.08  52     up         osd.33
>  34   hdd   3.73630  0.79999 3.7 TiB 2.9 TiB 2.8 TiB 449 MiB 7.7 GiB 877
> GiB 77.08 1.03  52     up         osd.34
>  35   hdd   3.73630  0.79999 3.7 TiB 2.3 TiB 2.2 TiB 133 MiB 7.0 GiB 1.4
> TiB 62.18 0.83  42     up         osd.35
>  36   hdd  14.65039  1.00000  15 TiB  11 TiB  11 TiB 544 MiB  26 GiB 4.0
> TiB 72.98 0.98 220     up         osd.36
>  37   hdd  14.65039  1.00000  15 TiB  11 TiB  11 TiB  11 GiB  38 GiB 3.6
> TiB 75.30 1.01 200     up         osd.37
>  38   hdd  14.65039  1.00000  15 TiB  11 TiB  11 TiB 1.2 GiB  28 GiB 3.3
> TiB 77.43 1.04 217     up         osd.38
>  -8        58.89636        -  59 TiB  47 TiB  46 TiB 3.9 GiB 116 GiB  12
> TiB 78.98 1.06   -            host s3db7
>  39   hdd   3.73630  1.00000 3.7 TiB 3.2 TiB 3.2 TiB  19 MiB 8.5 GiB 499
> GiB 86.96 1.17  43     up         osd.39
>  40   hdd   3.73630  1.00000 3.7 TiB 2.6 TiB 2.5 TiB 144 MiB 7.0 GiB 1.2
> TiB 68.33 0.92  39     up         osd.40
>  41   hdd   3.73630  1.00000 3.7 TiB 3.0 TiB 2.9 TiB 218 MiB 7.9 GiB 732
> GiB 80.86 1.08  64     up         osd.41
>  42   hdd   3.73630  1.00000 3.7 TiB 2.5 TiB 2.4 TiB 594 MiB 7.0 GiB 1.2
> TiB 67.97 0.91  50     up         osd.42
>  43   hdd  14.65039  1.00000  15 TiB  12 TiB  12 TiB 564 MiB  28 GiB 2.9
> TiB 80.32 1.08 213     up         osd.43
>  44   hdd  14.65039  1.00000  15 TiB  12 TiB  11 TiB 1.3 GiB  28 GiB 3.1
> TiB 78.59 1.05 198     up         osd.44
>  45   hdd  14.65039  1.00000  15 TiB  12 TiB  12 TiB 1.2 GiB  30 GiB 2.8
> TiB 81.05 1.09 214     up         osd.45
>  -9        51.28331        -  51 TiB  41 TiB  41 TiB 4.9 GiB 108 GiB  10
> TiB 79.75 1.07   -            host s3db8
>   8   hdd   7.32619  1.00000 7.3 TiB 5.8 TiB 5.8 TiB 472 MiB  15 GiB 1.5
> TiB 79.68 1.07  99     up         osd.8
>  16   hdd   7.32619  1.00000 7.3 TiB 5.9 TiB 5.8 TiB 785 MiB  15 GiB 1.4
> TiB 80.25 1.08  97     up         osd.16
>  31   hdd   7.32619  1.00000 7.3 TiB 5.5 TiB 5.5 TiB 438 MiB  14 GiB 1.8
> TiB 75.36 1.01  87     up         osd.31
>  52   hdd   7.32619  1.00000 7.3 TiB 5.7 TiB 5.7 TiB 844 MiB  15 GiB 1.6
> TiB 78.19 1.05 113     up         osd.52
>  53   hdd   7.32619  1.00000 7.3 TiB 6.2 TiB 6.1 TiB 792 MiB  18 GiB 1.1
> TiB 84.46 1.13 109     up         osd.53
>  54   hdd   7.32619  1.00000 7.3 TiB 5.6 TiB 5.6 TiB 959 MiB  15 GiB 1.7
> TiB 76.73 1.03 115     up         osd.54
>  55   hdd   7.32619  1.00000 7.3 TiB 6.1 TiB 6.1 TiB 699 MiB  16 GiB 1.2
> TiB 83.56 1.12 122     up         osd.55
> -10        51.28331        -  51 TiB  39 TiB  39 TiB 4.7 GiB 100 GiB  12
> TiB 76.05 1.02   -            host s3db9
>  56   hdd   7.32619  1.00000 7.3 TiB 5.2 TiB 5.2 TiB 840 MiB  13 GiB 2.1
> TiB 71.06 0.95 105     up         osd.56
>  57   hdd   7.32619  1.00000 7.3 TiB 6.1 TiB 6.0 TiB 1.0 GiB  16 GiB 1.2
> TiB 83.17 1.12 102     up         osd.57
>  58   hdd   7.32619  1.00000 7.3 TiB 6.0 TiB 5.9 TiB  43 MiB  15 GiB 1.4
> TiB 81.56 1.09 105     up         osd.58
>  59   hdd   7.32619  1.00000 7.3 TiB 5.9 TiB 5.9 TiB 429 MiB  15 GiB 1.4
> TiB 80.64 1.08  94     up         osd.59
>  60   hdd   7.32619  1.00000 7.3 TiB 5.4 TiB 5.3 TiB 226 MiB  14 GiB 2.0
> TiB 73.25 0.98 101     up         osd.60
>  61   hdd   7.32619  1.00000 7.3 TiB 4.8 TiB 4.8 TiB 1.1 GiB  12 GiB 2.5
> TiB 65.84 0.88 103     up         osd.61
>  62   hdd   7.32619  1.00000 7.3 TiB 5.6 TiB 5.6 TiB 1.0 GiB  15 GiB 1.7
> TiB 76.83 1.03 126     up         osd.62
>                        TOTAL 674 TiB 501 TiB 473 TiB  96 GiB 1.2 TiB 173
> TiB 74.57
> MIN/MAX VAR: 0.17/1.20  STDDEV: 10.25
>
>
>
> Am Sa., 13. März 2021 um 15:57 Uhr schrieb Dan van der Ster <
> dan@xxxxxxxxxxxxxx>:
>
>> No, increasing num PGs won't help substantially.
>>
>> Can you share the entire output of ceph osd df tree ?
>>
>> Did you already set
>>
>>   ceph config set mgr mgr/balancer/upmap_max_deviation 1
>>
>>
>> ??
>> And I recommend debug_mgr 4/5 so you can see some basic upmap balancer
>> logging.
>>
>> .. Dan
>>
>>
>>
>>
>>
>>
>> On Sat, Mar 13, 2021, 3:49 PM Boris Behrens <bb@xxxxxxxxx> wrote:
>>
>>> Hello people,
>>>
>>> I am still struggeling with the balancer
>>> (https://www.mail-archive.com/ceph-users@xxxxxxx/msg09124.html)
>>> Now I've read some more and might think that I do not have enough PGs.
>>> Currently I have 84OSDs and 1024PGs for the main pool (3008 total). I
>>> have the autoscaler enabled, but I doesn't tell me to increase the
>>> PGs.
>>>
>>> What do you think?
>>>
>>> --
>>> Die Selbsthilfegruppe "UTF-8-Probleme" trifft sich diesmal abweichend
>>> im groüen Saal.
>>> _______________________________________________
>>> ceph-users mailing list -- ceph-users@xxxxxxx
>>> To unsubscribe send an email to ceph-users-leave@xxxxxxx
>>>
>>
>
> --
> Die Selbsthilfegruppe "UTF-8-Probleme" trifft sich diesmal abweichend im
> groüen Saal.
>
_______________________________________________
ceph-users mailing list -- ceph-users@xxxxxxx
To unsubscribe send an email to ceph-users-leave@xxxxxxx




[Index of Archives]     [Information on CEPH]     [Linux Filesystem Development]     [Ceph Development]     [Ceph Large]     [Ceph Dev]     [Linux USB Development]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]     [xfs]


  Powered by Linux