Project

General

Profile

Actions

Bug #21331

closed

pg recovery priority inversion

Added by Sage Weil over 6 years ago. Updated over 6 years ago.

Status:
Resolved
Priority:
Urgent
Assignee:
Category:
-
Target version:
-
% Done:

0%

Source:
Tags:
Backport:
luminous
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Component(RADOS):
Pull request ID:
Crash signature (v1):
Crash signature (v2):

Description

32387/832388 n=10607 ec=1/1 lis/c 832387/827296 les/c/f 832388/827299/819533 832378/832387/832379) [92,18,26]/[92,18,16] r=0 lpr=832387 pi=[827296,832387)/3 bft=26 crt=832403'3721335 mlcod 0'0 active+recovery_wait+degraded+remapped m=26] execute_ctx 0:6c035d59:::10033c72719.00000000:head [create,setxattr parent (36
32387/832388 n=10607 ec=1/1 lis/c 832387/827296 les/c/f 832388/827299/819533 832378/832387/832379) [92,18,26]/[92,18,16] r=0 lpr=832387 pi=[827296,832387)/3 bft=26 crt=832403'3721335 mlcod 0'0 active+recovery_wait+degraded+remapped m=26] do_osd_op 0:6c035d59:::10033c72719.00000000:head [create,setxattr parent (369)
32387/832388 n=10607 ec=1/1 lis/c 832387/827296 les/c/f 832388/827299/819533 832378/832387/832379) [92,18,26]/[92,18,16] r=0 lpr=832387 pi=[827296,832387)/3 bft=26 crt=832403'3721335 mlcod 0'0 active+recovery_wait+degraded+remapped m=26] do_osd_op  create
32387/832388 n=10607 ec=1/1 lis/c 832387/827296 les/c/f 832388/827299/819533 832378/832387/832379) [92,18,26]/[92,18,16] r=0 lpr=832387 pi=[827296,832387)/3 bft=26 crt=832403'3721335 mlcod 0'0 active+recovery_wait+degraded+remapped m=26] do_osd_op  setxattr parent (369)
32387/832388 n=10607 ec=1/1 lis/c 832387/827296 les/c/f 832388/827299/819533 832378/832387/832379) [92,18,26]/[92,18,16] r=0 lpr=832387 pi=[827296,832387)/3 bft=26 crt=832403'3721335 mlcod 0'0 active+recovery_wait+degraded+remapped m=26] do_osd_op  setxattr layout (30)
32387/832388 n=10607 ec=1/1 lis/c 832387/827296 les/c/f 832388/827299/819533 832378/832387/832379) [92,18,26]/[92,18,16] r=0 lpr=832387 pi=[827296,832387)/3 bft=26 crt=832403'3721335 mlcod 0'0 active+recovery_wait+degraded+remapped m=26]  using newer snapc 1=[]
32387/832388 n=10607 ec=1/1 lis/c 832387/827296 les/c/f 832388/827299/819533 832378/832387/832379) [92,18,26]/[92,18,16] r=0 lpr=832387 pi=[827296,832387)/3 bft=26 crt=832403'3721335 mlcod 0'0 active+recovery_wait+degraded+remapped m=26]  set mtime to 2017-09-09 17:48:37.925036
32387/832388 n=10607 ec=1/1 lis/c 832387/827296 les/c/f 832388/827299/819533 832378/832387/832379) [92,18,26]/[92,18,16] r=0 lpr=832387 pi=[827296,832387)/3 bft=26 crt=832403'3721335 mlcod 0'0 active+recovery_wait+degraded+remapped m=26]  final snapset 1=[]:{} in 0:6c035d59:::10033c72719.00000000:head
32387/832388 n=10607 ec=1/1 lis/c 832387/827296 les/c/f 832388/827299/819533 832378/832387/832379) [92,18,26]/[92,18,16] r=0 lpr=832387 pi=[827296,832387)/3 bft=26 crt=832403'3721335 mlcod 0'0 active+recovery_wait+degraded+remapped m=26] new_repop rep_tid 2041 on osd_op(mds.0.308850:3887948 0.36 0:6c035d59:::10033c

notably does not trim in calc_trim_to. that's because mlcod = 0'0, and
<prE>
eversion_t limit = MIN(
min_last_complete_ondisk,
pg_log.get_can_rollback_to());
if (limit != eversion_t() &&
limit != pg_trim_to &&
pg_log.get_log().approx_size() > target) {

this pg has 54,000 entries! and is eating all the ram on the box.

Related issues 1 (0 open1 closed)

Related to Ceph - Bug #21761: ceph-osd consumes way too much memory during recoveryCan't reproduce10/11/2017

Actions
Actions #1

Updated by Sage Weil over 6 years ago

it looks lke peer_last_commit_ondisk for osd.26 isn't getting updated since it is not in acting (it's backfill target), and we aren't backfilling:

    "up": [
        92,
        18,
        26
    ],
    "acting": [
        92,
        18
    ],
    "backfill_targets": [
        "26" 
    ],
    "actingbackfill": [
        "18",
        "26",
        "92" 
    ],

and
    "state": "active+recovery_wait+undersized+degraded+remapped",

the peer info last_complete matches last_update,
        {
            "peer": "26",
            "pgid": "0.36",
            "last_update": "832429'3721336",
            "last_complete": "832429'3721336",

Actions #2

Updated by Sage Weil over 6 years ago

Actually, this isn't quite right.

The real problem is that the primary has an ancient last_complete, because it is in recovery_wait, and it is prioritized behind backfills that take forever.

Actions #3

Updated by Sage Weil over 6 years ago

Every 2.0s: ceph pg dump | sort -k 9 -r -n                                                                                                                                                                                                                                                                                                                                                                                                                                                               Sun Sep 10 00:39:55 2017

dumped all
sum 45550500 20 9651113 9742985 0 65275054173865 15707932 15707932
0  43745950 20 9154984 9157687 0 65271957084386 15350175 15350175
1   1803795  0  496129  585298 0      728765188   354060   354060
0.70b     10593                  0      245     10505       0 15725662580 95270    95270            active+recovery_wait+degraded+remapped 2017-09-09 19:59:22.472629  833302'3751117  833303:4957628       [80,0,69]         80      [80,69,61]             80  827049'3614209 2017-09-07 13:15:12.396269  827032'3547250 2017-09-04 19:29:05.528322
0.cc7     10770                  0      130     10764       0 15884739011 42096    42096            active+recovery_wait+degraded+remapped 2017-09-09 19:59:22.199903  833302'3441196  833303:4857777       [3,81,86]          3      [81,86,39]             81  827049'3330105 2017-09-07 01:37:13.997576  827049'3330105 2017-09-07 01:37:13.997576
0.943     10677                  0      177     10619       0 15862478339 38461    38461            active+recovery_wait+degraded+remapped 2017-09-09 19:59:27.757117  833302'3750408  833303:4761967       [74,2,12]         74      [74,12,39]             74  827049'3666300 2017-09-07 07:44:30.585682  827049'3666300 2017-09-07 07:44:30.585682
0.52e     10736                  0      135     10728       0 15950525996 35868    35868            active+recovery_wait+degraded+remapped 2017-09-09 19:59:26.259678  833302'3668952  833303:4804051      [86,118,0]         86      [86,118,1]             86  827045'3468140 2017-09-06 07:21:58.018293  827032'3460139 2017-09-04 17:40:00.982576
0.aaa     10759                  0      174     10715       0 16379572804 35086    35086            active+recovery_wait+degraded+remapped 2017-09-09 19:59:40.328664  833294'3643005  833303:5032293       [23,88,4]         23       [4,88,16]              4  827216'3555739 2017-09-07 17:36:34.743211  827044'3484483 2017-09-05 22:08:32.322834
0.92e     10680                  0      230     10606       0 16137890605 33170    33170            active+recovery_wait+degraded+remapped 2017-09-09 19:59:26.604688  833302'3367288  833302:4469626       [55,14,0]         55      [55,14,58]             55  827048'3268719 2017-09-06 16:23:48.155945  827031'3258244 2017-09-03 23:40:56.965737
0.4ce     10738                  0      243     10647       0 15928569742 31505    31505            active+recovery_wait+degraded+remapped 2017-09-09 19:59:38.363924  833302'3620237  833303:4680956      [37,12,23]         37      [37,12,21]             37  827049'3497915 2017-09-07 02:51:24.947311  827049'3497915 2017-09-07 02:51:24.947311
0.902     10616                  0      232     10525       0 15873498590 29006    29006            active+recovery_wait+degraded+remapped 2017-09-09 19:59:32.642511  833304'3393873  833304:4619034      [3,65,126]          3     [65,126,21]             65  827212'3270492 2017-09-07 15:36:53.468499  827031'3123446 2017-09-03 22:32:27.540699
0.c36     10793                  0      189     10748       0 16144758849 28721    28721            active+recovery_wait+degraded+remapped 2017-09-09 19:59:37.652769  833302'3333383  833303:4600623      [88,80,13]         88      [88,80,16]             88  827157'3231330 2017-09-07 15:11:29.356817  827044'3188787 2017-09-06 01:48:42.700501
0.a08     10609                  0      197     10529       0 15880133106 27685    27685            active+recovery_wait+degraded+remapped 2017-09-09 19:59:41.272426  833304'3373497  833304:4741143       [73,7,80]         73       [7,80,21]              7  827046'3310444 2017-09-06 12:02:47.620633  826961'3254016 2017-09-02 18:30:11.520031
0.45d     10663                  0        0     10185       0 16098153003 26356    26356            active+recovery_wait+degraded+remapped 2017-09-09 19:59:30.817526  833302'3317660  833302:4466927      [69,88,26]         69      [69,88,16]             69  827049'3176390 2017-09-06 21:39:57.724289  826974'3171907 2017-09-03 00:10:08.962798
0.9fd     10478                  0      281     10314       0 15550854041 25556    25556            active+recovery_wait+degraded+remapped 2017-09-09 19:59:24.585437  833294'3776002  833303:4784853      [81,14,73]         81      [81,14,39]             81  827049'3649706 2017-09-07 10:22:20.157062  827044'3622946 2017-09-06 03:40:46.832134
0.444     10773                  0      157     10746       0 16263547569 25341    25341            active+recovery_wait+degraded+remapped 2017-09-09 19:59:35.306642  833304'3915544  833304:4578979       [82,3,92]         82      [82,92,16]             82  827046'3735667 2017-09-06 09:51:40.545540  827032'3726546 2017-09-05 04:13:06.035978
0.b92     10527                  0      252     10422       0 15506757288 24602    24602            active+recovery_wait+degraded+remapped 2017-09-09 19:59:35.635577  833304'3407720  833304:4753268      [64,26,72]         64      [64,72,16]             64  827049'3326503 2017-09-07 06:01:47.559311  827044'3300304 2017-09-05 23:53:55.231309
0.61f     10644                  0      237     10561       0 16247319772 24293    24293            active+recovery_wait+degraded+remapped 2017-09-09 19:59:29.445424  833302'3824547  833302:4991671       [2,28,27]          2      [27,28,21]             27  827046'3728077 2017-09-06 14:47:42.234617  826949'3603169 2017-08-31 19:19:04.440645
0.d5a     10611                  0      212     10530       0 15713082461 23340    23340            active+recovery_wait+degraded+remapped 2017-09-09 19:59:37.386322  833302'3544887  833303:4967467       [13,57,4]         13       [4,57,21]              4  827254'3408556 2017-09-07 19:00:12.034747  826974'3299325 2017-09-03 03:07:44.889349
0.11b     10624                  0      152     10614       0 15868796961 22623    22623            active+recovery_wait+degraded+remapped 2017-09-09 19:59:34.847390  833302'3372294  833303:4458376      [18,72,73]         18      [18,72,63]             18  827049'3300020 2017-09-07 12:08:19.354375  827049'3300020 2017-09-07 12:08:19.354375
0.4f      10600                  0      119     10592       0 16058693765 22323    22323            active+recovery_wait+degraded+remapped 2017-09-09 19:59:22.150608  833302'3294469  833303:4707045      [36,26,87]         36      [36,87,58]             36  827049'3242468 2017-09-06 16:30:46.361715  826949'3145239 2017-08-31 22:24:59.386017
0.df9     10730                  0      181     10725       0 15905103992 22047    22047            active+recovery_wait+degraded+remapped 2017-09-09 19:59:20.828986  833302'3322425  833303:4716496       [0,68,65]          0      [68,65,39]             68  827049'3208000 2017-09-07 02:00:44.465320  827031'3154642 2017-09-04 10:06:02.669626
0.b5e     10585                  0      157     10560       0 15694156896 21939    21939            active+recovery_wait+degraded+remapped 2017-09-09 19:59:41.919490  833302'3317632  833303:4516003      [73,12,92]         73      [92,12,16]             92  827214'3246835 2017-09-07 16:46:45.341690  827044'3186624 2017-09-06 00:44:47.482618
0.832     10597                  0      226     10511       0 15810801929 21366    21366            active+recovery_wait+degraded+remapped 2017-09-09 19:59:26.262124  833302'3124730  833303:3844651      [37,74,13]         37      [37,74,61]             37  827049'3075918 2017-09-07 10:18:26.314923  827045'3053604 2017-09-06 05:06:08.366571
0.e5a     10708                  0      177     10701       0 16112566856 20980    20980            active+recovery_wait+degraded+remapped 2017-09-09 19:59:25.815887  833304'3172251  833304:4548005      [86,49,13]         86      [86,49,63]             86  827044'3070590 2017-09-06 01:01:22.269798  826961'3053074 2017-09-01 20:44:11.002632
0.cf8     10545                  0      221     10469       0 15677028730 20958    20958            active+recovery_wait+degraded+remapped 2017-09-09 19:59:36.912722  833302'3567639  833303:4985499      [87,13,14]         87      [87,14,16]             87  827044'3545021 2017-09-06 03:27:25.193698  827044'3545021 2017-09-06 03:27:25.193698
0.3b3     10465                  0      164     10432       0 15716760395 20734    20734            active+recovery_wait+degraded+remapped 2017-09-09 19:59:25.985937  833302'3692942  833303:4939972       [2,76,92]          2      [76,92,16]             76  827046'3561082 2017-09-06 11:01:58.331164  827031'3557366 2017-09-04 03:56:11.424728
0.f42     10631                  0      233     10544       0 15687177981 20235    20235            active+recovery_wait+degraded+remapped 2017-09-09 19:59:29.825357  833302'3323008  833303:4596713       [2,64,74]          2      [64,74,16]             64  827046'3210539 2017-09-06 11:42:08.829265  826961'3201394 2017-09-02 03:53:34.399678
0.ea3     10673                  0      234     10603       0 16262438334 20076    20076            active+recovery_wait+degraded+remapped 2017-09-09 19:59:37.912744  833302'3501374  833303:4229681        [65,2,7]         65       [65,7,21]             65  827049'3388083 2017-09-07 01:24:42.989196  827049'3388083 2017-09-07 01:24:42.989196
0.112     10535                  0        0      1372       0 15379251659 20024    20024            active+recovery_wait+degraded+remapped 2017-09-09 19:59:38.410230  833298'3256304  833303:4532353       [78,13,7]         78       [78,7,16]             78  827049'3179335 2017-09-07 12:39:59.424373  827031'3004530 2017-09-03 15:14:18.092065
0.a04     10634                  0      236     10536       0 16013799825 19990    19990            active+recovery_wait+degraded+remapped 2017-09-09 19:59:26.467989  833302'3602740  833302:4995717      [26,12,72]         26      [72,12,39]             72  827046'3433525 2017-09-06 12:27:32.299217  826948'3306706 2017-08-31 11:37:21.026743
0.569     10880                  0      255     10777       0 16630657275 19757    19757            active+recovery_wait+degraded+remapped 2017-09-09 19:59:27.125163  833302'3836654  833303:5015974      [0,127,38]          0     [38,127,16]             38  827049'3778516 2017-09-07 10:10:02.308094  827049'3778516 2017-09-07 10:10:02.308094
0.e02     10633                  0      170     10632       0 15800551075 19681    19681            active+recovery_wait+degraded+remapped 2017-09-09 19:59:19.974019  833296'3887885  833303:5170175      [26,78,37]         26      [78,37,63]             78  827049'3808533 2017-09-06 19:48:34.608093  827049'3808533 2017-09-06 19:48:34.608093
0.55f     10551                  0      204     10472       0 16005710365 19677    19677            active+recovery_wait+degraded+remapped 2017-09-09 19:59:34.492007  833302'3843720  833303:4920786      [10,23,72]         10      [10,72,21]             10  827049'3740823 2017-09-07 12:16:06.374705  827044'3700959 2017-09-06 02:18:07.146949
0.66e     10676                  0      155     10650       0 16315958135 19650    19650            active+recovery_wait+degraded+remapped 2017-09-09 19:59:36.598138  833302'3345479  833303:4722157       [44,2,87]         44      [44,87,21]             44  827049'3217768 2017-09-06 21:41:50.211526  827032'3202921 2017-09-04 11:38:23.217519
0.516     10704                  0      144     10691       0 15942599865 19527    19527            active+recovery_wait+degraded+remapped 2017-09-09 19:59:36.784746  833302'3890669  833303:5016703      [13,12,85]         13      [12,85,21]             12  827045'3731023 2017-09-06 07:37:13.591157  827032'3721082 2017-09-04 21:58:47.408326
0.a99     10475                  0      192     10402       0 15482148590 18601    18601            active+recovery_wait+degraded+remapped 2017-09-09 19:59:39.394797  833298'3576867  833303:4644812      [75,73,22]         75      [75,22,21]             75  827049'3533656 2017-09-07 03:48:01.620665  826961'3489050 2017-09-02 04:54:50.212676
0.c7e     10630                  0      140     10624       0 15863216162 18442    18442            active+recovery_wait+degraded+remapped 2017-09-09 19:59:27.229278  833298'3442566  833303:4615082       [67,8,73]         67       [67,8,63]             67  827049'3393659 2017-09-07 11:22:59.195776  827031'3278657 2017-09-03 10:27:19.049459
0.29d     10577                  0      167     10572       0 15486023766 18407    18407            active+recovery_wait+degraded+remapped 2017-09-09 19:59:36.079278  833302'3018278  833303:4265821       [85,83,3]         85      [85,83,21]             85  827045'2913307 2017-09-06 07:22:59.680943  826949'2818578 2017-09-01 00:33:11.980807
0.bf      10814                  0      245     10724       0 16134801877 18365    18365            active+recovery_wait+degraded+remapped 2017-09-09 19:59:27.780406  833302'3758150  833303:4810283      [54,13,12]         54      [54,12,58]             54  827046'3672809 2017-09-06 12:38:17.275954  827046'3672809 2017-09-06 12:38:17.275954
0.70a     10879                  0      210     10802       0 16304665542 17734    17734            active+recovery_wait+degraded+remapped 2017-09-09 19:59:32.664614  833294'3585869  833303:4530756       [123,1,0]        123      [123,1,21]            123  827049'3480220 2017-09-07 09:54:24.770979  827049'3480220 2017-09-07 09:54:24.770979
0.179     10848                  0      170     10825       0 16239330579 17675    17675            active+recovery_wait+degraded+remapped 2017-09-09 19:59:30.059594  833302'3200974  833302:4464954       [90,26,9]         90       [90,9,61]             90  827046'3065091 2017-09-06 13:10:14.612719  826947'2921834 2017-08-31 11:07:04.193592
0.1f5     10636                  0      137     10627       0 16341173209 17501    17501            active+recovery_wait+degraded+remapped 2017-09-09 19:59:27.726357  833282'3917749  833303:5312050      [73,80,27]         73      [80,27,63]             80  827049'3805168 2017-09-06 20:48:53.216754  826948'3637330 2017-08-31 18:41:03.572123
0.922     10787                  0      255     10684       0 16213696316 17301    17301            active+recovery_wait+degraded+remapped 2017-09-09 19:59:31.147799  833302'3709168  833303:4569542      [18,73,54]         18      [18,54,61]             18  827049'3665744 2017-09-06 19:07:37.225860  827049'3665744 2017-09-06 19:07:37.225860
0.790     10717                  0      176     10695       0 16085501398 17222    17222            active+recovery_wait+degraded+remapped 2017-09-09 19:59:24.549833  833302'3636199  833302:4638291       [3,91,55]          3      [55,91,21]             55  827049'3551647 2017-09-07 06:30:50.551217  827049'3551647 2017-09-07 06:30:50.551217
0.f00     10654                  0      219     10582       0 15852930963 17012    17012            active+recovery_wait+degraded+remapped 2017-09-09 19:59:28.213869  833300'3528802  833303:4856770        [74,2,7]         74       [74,7,39]             74  827049'3446844 2017-09-06 23:51:09.563708  827032'3441163 2017-09-04 11:47:53.146352
0.13a     10805                  0      130     10795       0 16275357737 16994    16994            active+recovery_wait+degraded+remapped 2017-09-09 19:59:32.361925  833302'3553539  833303:4748318        [0,71,9]          0       [71,9,63]             71  827216'3521409 2017-09-07 17:42:42.457373  826974'3346887 2017-09-03 05:04:59.572873
0.5fc     10734                  0      249     10642       0 16598177230 16904    16904            active+recovery_wait+degraded+remapped 2017-09-09 19:59:26.792549  833303'3613319  833303:4621328       [69,86,2]         69      [69,86,16]             69  827049'3554170 2017-09-07 10:54:43.536665  826961'3462137 2017-09-01 20:00:31.110685
0.38d     10652                  0      179     10618       0 15883070617 16667    16667            active+recovery_wait+degraded+remapped 2017-09-09 19:59:34.956318  833302'3573286  833303:4730071       [86,66,2]         86      [86,66,21]             86  827044'3477117 2017-09-05 21:51:29.628451  827032'3467067 2017-09-04 16:55:14.044461
0.762     10751                  0      217     10670       0 15833396341 15767    15767            active+recovery_wait+degraded+remapped 2017-09-09 19:59:29.363695  833302'3643403  833303:4439600      [44,26,18]         44      [44,18,58]             44  827046'3575758 2017-09-06 13:36:53.894145  827037'3573783 2017-09-05 06:23:46.659999
0.17f     10640                  0      209     10552       0 15845344084 15739    15739            active+recovery_wait+degraded+remapped 2017-09-09 19:59:28.337027  833294'3520864  833303:4810800      [57,26,79]         57      [57,79,16]             57  827046'3451914 2017-09-06 11:00:05.305222  826905'3324791 2017-08-31 05:01:13.747840
0.722     10519                  0      235     10425       0 15594727416 15289    15289            active+recovery_wait+degraded+remapped 2017-09-09 19:59:25.893809  833296'3366065  833303:4758038      [37,13,52]         37      [37,52,61]             37  827049'3282519 2017-09-07 02:39:56.577494  827044'3277451 2017-09-05 21:45:17.200006
0.9c0     10594                  0      149     10584       0 15881663550 15249    15249            active+recovery_wait+degraded+remapped 2017-09-09 19:59:34.485062  833302'3569622  833303:4734351       [85,74,0]         85      [85,74,21]             85  827046'3463055 2017-09-06 10:16:38.577063  827046'3463055 2017-09-06 10:16:38.577063
0.a5e     10798                  0      246     10678       0 16165122510 14816    14816            active+recovery_wait+degraded+remapped 2017-09-09 19:59:24.740378  833290'3535234  833303:4919012      [13,47,27]         13      [47,27,39]             47  827046'3484379 2017-09-06 12:01:16.783422  827046'3484379 2017-09-06 12:01:16.783422
0.eee     10596                  0      235     10506       0 16044617744 14188    14188            active+recovery_wait+degraded+remapped 2017-09-09 19:59:28.984580  833302'3438271  833303:4393250      [24,27,73]         24      [24,27,16]             24  827046'3349187 2017-09-06 15:30:02.160142  827037'3324448 2017-09-05 12:37:58.271070
0.ea2     10806                  0      131     10800       0 16255613624 13630    13630            active+recovery_wait+degraded+remapped 2017-09-09 19:59:34.979282  833298'3550497  833303:4643900      [10,18,73]         10      [10,18,63]             10  827049'3486677 2017-09-06 16:56:31.368640  826941'3346779 2017-08-31 09:07:35.080404
0.d4a     10969                  0      221     10920       0 16310759119 13395    13395            active+recovery_wait+degraded+remapped 2017-09-09 19:59:37.260946  833302'3572684  833303:4697029       [85,2,37]         85      [85,37,16]             85  827046'3483009 2017-09-06 13:13:31.522250  827046'3483009 2017-09-06 13:13:31.522250
0.326     10734                  0      162     10712       0 16241669113 12637    12637            active+recovery_wait+degraded+remapped 2017-09-09 19:59:29.688602  833302'3688707  833303:4927076       [86,24,0]         86      [86,24,61]             86  827046'3611428 2017-09-06 12:22:09.884754  826205'3422424 2017-08-30 15:31:11.760783
0.527     10727                  0      246     10635       0 16079562562 12636    12636            active+recovery_wait+degraded+remapped 2017-09-09 19:59:39.418691  833302'3708274  833302:5335339      [55,10,13]         55      [55,10,21]             55  827046'3571290 2017-09-06 14:43:25.814125  826340'3440177 2017-08-30 17:55:12.476191
0.174     10894                  0      234     10821       0 16427577906 12523    12523            active+recovery_wait+degraded+remapped 2017-09-09 19:59:20.151760  833302'3335191  833303:4422773      [26,80,49]         26      [80,49,58]             80  827049'3304622 2017-09-06 21:31:16.897195  826948'3224052 2017-08-31 15:51:10.061613
0.214     10754                  0      344     10551       0 15848213511 12345    12345            active+recovery_wait+degraded+remapped 2017-09-09 19:59:27.059585  833302'3288352  833303:4535546     [18,126,73]         18     [18,126,39]             18  827049'3241852 2017-09-07 04:13:51.608020  827049'3241852 2017-09-07 04:13:51.608020
0.921     10727                  0      225     10635       0 15970415790 10926    10926            active+recovery_wait+degraded+remapped 2017-09-09 19:59:30.648425  833302'4050011  833302:5329965     [34,126,23]         34     [34,126,16]             34  827049'3996172 2017-09-07 10:32:54.567461  827032'3942656 2017-09-04 13:47:18.040718
0.a65     10750                  0      138     10745       0 16534988186 10908    10908            active+recovery_wait+degraded+remapped 2017-09-09 19:59:26.186671  833304'3668205  833304:4861656      [92,23,28]         92      [92,28,39]             92  827046'3536746 2017-09-06 14:00:55.992456  827032'3531584 2017-09-05 04:00:02.712845
0.240     10804                  0      186     10791       0 16146428696 10203    10203            active+recovery_wait+degraded+remapped 2017-09-09 19:59:30.576625  833302'3681339  833302:4681889      [26,91,66]         26      [66,91,21]             66  827045'3488797 2017-09-06 04:32:28.044557  826923'3404384 2017-08-31 07:02:10.438613
0.354     10702                  0      173     10675       0 15431796310 10182    10182            active+recovery_wait+degraded+remapped 2017-09-09 19:59:31.127433  833302'3237879  833302:4256393      [66,92,23]         66      [66,92,16]             66  827046'3174218 2017-09-06 15:48:06.913172  827031'3158225 2017-09-04 09:34:28.495865
0.512     10620                  0      237     10517       0 16070207959 10154    10154            active+recovery_wait+degraded+remapped 2017-09-09 19:59:41.085735  833294'4189589  833303:5430286        [1,0,86]          1       [1,86,16]              1  827049'4027505 2017-09-06 17:15:16.166582  826441'3836069 2017-08-30 20:08:27.699655
0.b7b     10790                  0    10786     10712       0 15832506785 10148    10148 active+recovery_wait+undersized+degraded+remapped 2017-09-09 19:59:36.949896  833304'3473183  833304:4634855        [1,0,19]          1          [1,21]              1  827049'3393728 2017-09-06 16:47:30.128239  827037'3391731 2017-09-05 11:49:00.446243
0.e10     10670                  0    10721         0       0 15860765136 10100    10100 active+undersized+degraded+remapped+backfill_wait 2017-09-09 19:59:23.918310  833302'3399041  833303:4398319      [24,19,64]         24         [24,64]             24  827214'3314112 2017-09-07 16:45:25.530227  826949'3167695 2017-08-31 21:08:38.250720
0.9b2     10859                  0    10929         0       0 15975049643 10100    10100 active+undersized+degraded+remapped+backfill_wait 2017-09-09 19:59:26.502221  833302'3656929  833303:4484313       [1,15,34]          1          [1,34]              1  827049'3591687 2017-09-06 23:45:13.672278  826974'3499516 2017-09-03 02:56:21.554506
0.550     10684                  0      123     10684       0 15734743347 10100    10100            active+degraded+remapped+backfill_wait 2017-09-09 22:01:26.947805  833302'3523235  833303:4451255     [123,69,26]        123     [123,69,63]            123  827049'3437773 2017-09-07 13:08:27.619006  827049'3437773 2017-09-07 13:08:27.619006
0.124     10669                  0    10749         0       0 16185101353 10100    10100 active+undersized+degraded+remapped+backfill_wait 2017-09-09 19:59:23.124036  833302'3703924  833303:4816640      [37,76,26]         37         [37,76]             37  827049'3606032 2017-09-07 08:08:32.135760  827032'3558477 2017-09-04 18:17:28.262969
0.e3a     10739                  0    10793     10739       0 16253620047 10099    10099 active+undersized+degraded+remapped+backfill_wait 2017-09-09 19:59:28.194934  833302'3736309  833302:4890298       [2,15,72]          2         [72,63]             72  827049'3592574 2017-09-06 20:16:43.894841  826949'3478916 2017-08-31 20:37:52.811838
0.887     10414                  0     2816         0       0 15614065866 10099    10099 active+undersized+degraded+remapped+backfill_wait 2017-09-09 19:59:26.192917  833302'3914305  833303:5296547       [92,22,6]         92         [92,22]             92  827045'3848572 2017-09-06 07:49:20.592645  827045'3848572 2017-09-06 07:49:20.592645
0.e47     10499                  0      110     10499       0 15782413239 10098    10098                     active+remapped+backfill_wait 2017-09-09 19:59:22.137933  833302'3449783  833303:4553697      [65,126,3]         65     [65,126,61]             65  827046'3309913 2017-09-06 12:53:04.498618  827031'3226276 2017-09-03 22:22:24.853794
0.e35     10706                  0    10788         0       0 15780722548 10098    10098 active+undersized+degraded+remapped+backfill_wait 2017-09-09 19:59:20.629337  833302'3439415  833303:4670932     [17,80,127]         17        [80,127]             80  827049'3361408 2017-09-07 10:33:03.925504  827044'3335615 2017-09-06 02:02:56.409389
0.b65     10757                  0    10822         0       0 15848308168 10098    10098 active+undersized+degraded+remapped+backfill_wait 2017-09-09 19:59:26.081532  833288'3528799  833303:4551462      [15,36,12]         15         [36,12]             36  827049'3443149 2017-09-07 08:43:15.840851  827044'3400136 2017-09-05 23:10:55.214032

column 9 is the pg log length. note that the backfill_wait ones have big logs.

workaround to make them recover and trim is

ceph osd set nobackfill
ceph tell osd.* config set osd_max_backfills 500

although i think if you set osd_max_backfills back to 1 and unset nobackfill they will all backfill in parallel?
Actions #4

Updated by Sage Weil over 6 years ago

  • Subject changed from pg log not getting trimmed while degraded to pg recovery priority inversion
  • Priority changed from Immediate to Urgent
Actions #5

Updated by Sage Weil over 6 years ago

  • Status changed from 12 to Resolved
  • Backport set to luminous
Actions #6

Updated by Sage Weil over 6 years ago

  • Related to Bug #21761: ceph-osd consumes way too much memory during recovery added
Actions

Also available in: Atom PDF