Re: Accidentally Remove OSDs

[FaHui Lin <fahui.lin@xxxxxxxxxx>]

Hi, thank you for your response.

Well, I've not only taken out but also totally removed the both OSDs (by "ceph osd rm" and delete everything in /var/lib/ceph/osd/<related OSDs>) of that pg (and similar to all other stale pgs.)

The main problem I have is those stale pgs (miss all OSDs I've removed) not merely make ceph health warning, but other machine cannot mount the ceph rbd as well.

Here's the full crush map.  The OSDs I removed were osd.5~19.

# begin crush map
tunable choose_local_tries 0
tunable choose_local_fallback_tries 0
tunable choose_total_tries 500

# devices
device 0 osd.0
device 1 device1
device 2 osd.2
device 3 osd.3
device 4 osd.4
device 5 device5
device 6 device6
device 7 device7
device 8 device8
device 9 device9
device 10 device10
device 11 device11
device 12 device12
device 13 device13
device 14 device14
device 15 device15
device 16 device16
device 17 device17
device 18 device18
device 19 device19
device 20 osd.20
device 21 osd.21
device 22 osd.22
device 23 osd.23
device 24 osd.24
device 25 osd.25
device 26 osd.26
device 27 osd.27

# types
type 0 osd
type 1 host
type 2 rack
type 3 row
type 4 room
type 5 datacenter
type 6 root

# buckets
host XX-ceph01 {
        id -2           # do not change unnecessarily
        # weight 160.040
        alg straw
        hash 0  # rjenkins1
        item osd.0 weight 40.010
        item osd.2 weight 40.010
        item osd.3 weight 40.010
        item osd.4 weight 40.010
}
host XX-ceph02 {
        id -3           # do not change unnecessarily
        # weight 320.160
        alg straw
        hash 0  # rjenkins1
        item osd.20 weight 40.020
        item osd.21 weight 40.020
        item osd.22 weight 40.020
        item osd.23 weight 40.020
        item osd.24 weight 40.020
        item osd.25 weight 40.020
        item osd.26 weight 40.020
        item osd.27 weight 40.020
}
root default {
        id -1           # do not change unnecessarily
        # weight 480.200
        alg straw
        hash 0  # rjenkins1
        item XX-ceph01 weight 160.040
        item XX-ceph02 weight 320.160
}

# rules
rule data {
        ruleset 0
        type replicated
        min_size 1
        max_size 10
        step take default
        step chooseleaf firstn 0 type host
        step emit
}
rule metadata {
        ruleset 1
        type replicated
        min_size 1
        max_size 10
        step take default
        step chooseleaf firstn 0 type host
        step emit
}
rule rbd {
        ruleset 2
        type replicated
        min_size 1
        max_size 10
        step take default
        step chooseleaf firstn 0 type host
        step emit
}

# end crush map

List of some stale pgs:

pg_stat objects mip     degr    misp    unf     bytes   log     disklog state   state_stamp     v       reported        up      up_primary      acting  acting_primary  last_scrub      scrub_stamp     last_deep_scrub deep_scrub_stamp
17.c6   0       0       0       0       0       0       0       0       stale+active+clean      2015-04-20 09:16:09.358613      0'0     2706:216        [19,13] 19      [19,13] 19      0'0     2015-04-16 02:29:34.882038
      0'0     2015-04-16 02:29:34.882038
17.c7   0       0       0       0       0       0       0       0       stale+active+clean      2015-04-20 09:16:28.304621      0'0     2718:262        [15,18] 15      [15,18] 15      0'0     2015-04-20 09:15:39.363310
      0'0     2015-04-20 09:15:39.363310
17.c1   0       0       0       0       0       0       0       0       stale+active+clean      2015-04-20 09:16:01.073681      0'0     2706:199        [19,16] 19      [19,16] 19      0'0     2015-04-15 12:37:11.741251
      0'0     2015-04-15 12:37:11.741251
17.de   0       0       0       0       0       0       0       0       stale+active+undersized+degraded        2015-04-20 23:41:29.436796      0'0     2718:267        [15]    15      [15]    15      0'0     2015-04-13 07:56:01.760824      0'0     2015-04-13 07:56:01.760824
17.da   0       0       0       0       0       0       0       0       stale+active+undersized+degraded        2015-04-20 23:41:50.001087      0'0     2718:232        [14]    14      [14]    14      0'0     2015-04-19 15:45:53.304596      0'0     2015-04-19 15:45:53.304596
17.d9   0       0       0       0       0       0       0       0       stale+active+undersized+degraded        2015-04-20 23:41:29.472983      0'0     2718:270        [14]    14      [14]    14      0'0     2015-04-16 01:55:44.183550      0'0     2015-04-16 01:55:44.183550
17.d7   0       0       0       0       0       0       0       0       stale+active+undersized+degraded        2015-04-20 23:41:53.839134      0'0     2718:68 [17]    17      [17]    17      0'0     2015-04-16 00:06:27.998210      0'0     2015-04-16 00:06:27.998210
17.d5   0       0       0       0       0       0       0       0       stale+active+clean      2015-04-20 09:16:28.311352      0'0     2718:226        [18,17] 18      [18,17] 18      0'0     2015-04-15 20:52:33.372369
      0'0     2015-04-15 20:52:33.372369
17.d0   0       0       0       0       0       0       0       0       stale+active+clean      2015-04-20 09:16:24.850188      0'0     2718:213        [15,12] 15      [15,12] 15      0'0     2015-04-19 15:40:32.215234
      0'0     2015-04-19 15:40:32.215234
17.d1   0       0       0       0       0       0       0       0       stale+active+clean      2015-04-20 09:16:24.849996      0'0     2718:227        [15,12] 15      [15,12] 15      0'0     2015-04-15 19:03:38.137147
      0'0     2015-04-15 19:03:38.137147
17.ae   0       0       0       0       0       0       0       0       stale+active+clean      2015-04-20 09:16:28.310506      0'0     2718:231        [18,12] 18      [18,12] 18      0'0     2015-04-16 02:23:35.031329
      0'0     2015-04-16 02:23:35.031329
17.ac   0       0       0       0       0       0       0       0       stale+active+undersized+degraded        2015-04-20 23:41:50.002406      0'0     2718:66 [12]    12      [12]    12      0'0     2015-04-16 02:23:33.023476      0'0     2015-04-16 02:23:33.023476
17.aa   0       0       0       0       0       0       0       0       stale+active+clean      2015-04-20 09:16:25.983034      0'0     2718:213        [15,14] 15      [15,14] 15      0'0     2015-04-19 15:32:38.896039
      0'0     2015-04-19 15:32:38.896039
17.ab   0       0       0       0       0       0       0       0       stale+active+clean      2015-04-20 09:16:24.836133      0'0     2718:260        [12,17] 12      [12,17] 12      0'0     2015-04-19 15:32:44.905707
      0'0     2015-04-19 15:32:44.905707
17.a8   0       0       0       0       0       0       0       0       stale+active+clean      2015-04-20 09:16:09.361319      0'0     2706:212        [19,13] 19      [19,13] 19      0'0     2015-04-16 02:23:32.026015
      0'0     2015-04-16 02:23:32.026015
17.a6   0       0       0       0       0       0       0       0       stale+active+undersized+degraded        2015-04-20 23:41:50.002804      0'0     2718:96 [18]    18      [18]    18      0'0     2015-04-20 14:02:29.334181      0'0     2015-04-20 14:02:29.334181
17.a4   0       0       0       0       0       0       0       0       stale+active+clean      2015-04-20 09:16:28.310707      0'0     2718:232        [18,17] 18      [18,17] 18      0'0     2015-04-16 02:22:12.018136
      0'0     2015-04-16 02:22:12.018136
17.a2   0       0       0       0       0       0       0       0       stale+active+clean      2015-04-20 09:16:11.624952      0'0     2718:200        [15,17] 15      [15,17] 15      0'0     2015-04-15 10:42:37.880699
      0'0     2015-04-15 10:42:37.880699
17.a0   0       0       0       0       0       0       0       0       stale+active+undersized+degraded        2015-04-20 23:41:29.469600      0'0     2718:66 [18]    18      [18]    18      0'0     2015-04-16 02:22:08.992748      0'0     2015-04-16 02:22:08.992748

OSDs of those pgs (either primary or secondary) are totally gone, and I cannot find a way to repair them.

I've had another machince of new drive partitions, and I tried to re-create OSDs I had removed on it, but that would be osd.28, 29, etc. That's why I wondered how to change ID number of an OSD.

Regardless of the data loss (which I think it's already happened), I'd like to make the ceph service normal asap.
Is there anyway to deal with those stale pgs? (such as to recreate the OSDs they need, or to inject exsisting OSDs to those pgs, or even to kill those pgs?)
And since I'm not experienced, I may need more concrete comments (i.e. approach with ceph commands). Many thaks for your help.

Best Regards,
FaHui

Robert LeBlanc 於 2015/4/23 下午 10:53 寫道:

A full CRUSH dump would be helpful, as well as knowing which OSDs you took out. If you didn't take 17 out as well as 15, then you might be OK. If the OSDs still show up in your CRUSH, then try and remove them from the CRSH map with 'ceph osd crush rm osd.15'.

If you took out both OSDs, you will need to use some of the recovery tools. I believe the procedure is roughly, mount the drive in another box, extract the PGs needed, then shut down the primary OSD for that PG, inject the PG into the OSD, then start it up and it should replicate. I haven't done it myself (probably something I should do in case I ever run into the problem).

On Thu, Apr 23, 2015 at 2:00 AM, FaHui Lin <fahui.lin@xxxxxxxxxx> wrote:

Dear Ceph experts,

I'm a very new Ceph user. I made a blunder that I removed some OSDs (and all files in the related directories) before Ceph finished rebalancing datas and migrating pgs.

Not to mention the data loss, I meet the problem that:

1) There are always stale pgs showing in ceph status (with heath warning). Say one of the stale pg 17.a2:

# ceph -v
ceph version 0.87.1 (283c2e7cfa2457799f534744d7d549f83ea1335e)

# ceph -s
    cluster 3f81b47e-fb15-4fbb-9fee-0b1986dfd7ea
     health HEALTH_WARN 203 pgs degraded; 366 pgs stale; 203 pgs stuck degraded; 366 pgs stuck stale; 203 pgs stuck unclean; 203 pgs stuck undersized; 203 pgs undersized; 154 requests are blocked > 32 sec; recovery 153738/18991802 objects degraded (0.809%)
     monmap e1: 1 mons at {...=...:6789/0}, election epoch 1, quorum 0 tw-ceph01
     osdmap e3697: 12 osds: 12 up, 12 in
      pgmap v21296531: 1156 pgs, 18 pools, 36929 GB data, 9273 kobjects
            72068 GB used, 409 TB / 480 TB avail
            153738/18991802 objects degraded (0.809%)
                 163 stale+active+clean
                 786 active+clean
                 203 stale+active+undersized+degraded
                   4 active+clean+scrubbing+deep


# ceph pg dump_stuck stale | grep 17.a2
17.a2   0       0       0       0       0       0       0       0       stale+active+clean      2015-04-20 09:16:11.624952     0'0     2718:200        [15,17] 15      [15,17] 15      0'0     2015-04-15 10:42:37.880699    0'0      2015-04-15 10:42:37.880699

# ceph pg repair 17.a2
Error EAGAIN: pg 17.a2 primary osd.15 not up

# ceph pg scrub 17.a2
Error EAGAIN: pg 17.a2 primary osd.15 not up

# ceph pg map 17.a2
osdmap e3695 pg 17.a2 (17.a2) -> up [27,3] acting [27,3]

where osd.15 had already been removed. It seems to map to the existing OSDs ([27, 3]).
Can this pg finally get recovered by changing to the existing OSDs? If not, how can I do about this kind of stale pg?

2) I tried to solve the problem above by creating OSDs back but failed. The reason was I cannot create an OSD with the same ID to that I removed, say osd.15 (or change the id of an OSD).
Is there any way to change the id of an OSD? (By the way, I'm suprised that this issue can hardly be found on the internet.)

3) I tried another thing: to dump the crushmap and remove everything (including devices and buckets sections) related to the OSDs I removed. However, after I set the crushmap and dumped it out again, I found the OSDs's line still appear in the devices section (not in the buckets section though), such as:

# devices
device 0 osd.0
device 2 osd.2
device 3 osd.3
device 4 osd.4
device 5 device5
...
device 14 device14
device 15 device15

Is there anyway to remove them? Does it matters when I want to add new OSDs?

Please inform me if you have any comments. Thank you.

Best Regards,
FaHui

_______________________________________________
ceph-users mailing list
ceph-users@xxxxxxxxxxxxxx
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com

_______________________________________________
ceph-users mailing list
ceph-users@xxxxxxxxxxxxxx
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com