On Thu, Sep 27, 2012 at 3:52 AM, hemant surale <hemant.surale@xxxxxxxxx> wrote:
> Sir,
> I have upgraded my cluster to Ceph v0.48. and cluster is fine (except
> gceph not working) .
>
> How can I direct my data inject to specific osds ?
> i tried to edit crushmap & tired to specify ruleset accordingly , But
> its not working ( I did same for Ceph v0.36 it worked )
>
> -----------------------------------------------------------------------//crushmap-----------------------------------
> # begin crush map
>
> # devices
> device 0 osd.0
> device 1 osd.1
>
> # types
> type 0 osd
> type 1 host
> type 2 rack
> type 3 row
> type 4 room
> type 5 datacenter
> type 6 pool
> type 7 ghost
>
> # buckets
> host atish-virtual-machine {
> id -2 # do not change unnecessarily
> # weight 1.000
> alg straw
> hash 0 # rjenkins1
> item osd.0 weight 1.000
> }
> ghost bob-virtual-machine {
> id -4 # do not change unnecessarily
> # weight 1.000
> alg straw
> hash 0 # rjenkins1
> item osd.1 weight 1.000
> }
> rack unknownrack {
> id -3 # do not change unnecessarily
> # weight 2.000
> alg straw
> hash 0 # rjenkins1
> item atish-virtual-machine weight 1.000
> item bob-virtual-machine weight 1.000
> }
> pool default {
> id -1 # do not change unnecessarily
> # weight 2.000
> alg straw
> hash 0 # rjenkins1
> item unknownrack weight 2.000
> }
>
> # 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
> }
> rule newbyh {
> ruleset 3
> type replicated
> min_size 1
> max_size 10
> step take default
> step chooseleaf firstn 0 type ghost
> step emit
> }
>
>
> # end crush map
> ---------------------------------------------//EOF-------------------------------------------------------------------------
> I've created new pool "new" set its replication factor to 1 & also set
> crush_ruleset property to 3 .I was expecting ceph to select osd.1 i.e.
> bob-virtual-machine but some(Obj1,Obj2) Objects are placed into osd.0
> (atish-V-M) and some(Obj3) was placed in osd.1. Does my crushmap have
> errors ? Or I am missing the direction ?
To fix this specific error, you'll want to change it from
step take default
step chooseleaf firstn 0 type ghost
to
step take bob-virtual-machine
step chooseleaf firstn 0 type osd
But this isn't a good plan for doing data layout in general — it seems
like Ceph is not a fit for your needs if you need to specify that
specific data goes to specific OSDs.
-Greg
>
>
> -
> Hemant Surale
>
>
>
>
>
> On Thu, Sep 27, 2012 at 10:56 AM, hemant surale <hemant.surale@xxxxxxxxx> wrote:
>> Sure Sir ! :)
>>
>>
>> On Thu, Sep 27, 2012 at 10:50 AM, Dan Mick <dan.mick@xxxxxxxxxxx> wrote:
>>> OK. Let us know if osd map ends up working out for you.
>>>
>>>
>>> On 09/26/2012 10:18 PM, hemant surale wrote:
>>>>
>>>> Sir,
>>>> I am now upgrading to latest version of ceph. I was using 0.36
>>>> because it was running fine on 3 node cluster & i was really afraid of
>>>> problem could pop up , But learning is something that consists of
>>>> problem solving . So np.
>>>>
>>>> -
>>>> Hemant Surale.
>>>>
>>>> On Thu, Sep 27, 2012 at 10:38 AM, Dan Mick <dan.mick@xxxxxxxxxxx> wrote:
>>>>>
>>>>> 0.36?? That is *very* old. Is there a good reason not to upgrade?
>>>>>
>>>>>
>>>>> On 09/26/2012 07:31 PM, hemant surale wrote:
>>>>>>
>>>>>>
>>>>>> Sir,
>>>>>> I am using Ceph v0.36 . May be thats the reason after executing cmd
>>>>>> told by you execution goes like this
>>>>>>
>>>>>>
>>>>>> ---------------------------------------------
>>>>>> root@third-virtual-machine:~# ceph osd map newbyh Obj1
>>>>>> 2012-09-26 12:17:49.900514 mon <- [osd,map,newbyh,Obj1]
>>>>>> 2012-09-26 12:17:49.900977 mon0 -> 'unknown command map' (-22)
>>>>>> ----------------------------------------------
>>>>>>
>>>>>>
>>>>>>
>>>>>> -
>>>>>> Hemant Surale.
>>>>>>
>>>>>>
>>>>>> On Thu, Sep 27, 2012 at 3:23 AM, Dan Mick <dan.mick@xxxxxxxxxxx> wrote:
>>>>>>>
>>>>>>>
>>>>>>> Ah, yeah, that assumption would be a problem.
>>>>>>>
>>>>>>> So, Hemant, does
>>>>>>> ceph osd dump <poolname> <objectname>
>>>>>>>
>>>>>>> show you information that makes sense?
>>>>>>>
>>>>>>>
>>>>>>> On 09/26/2012 08:21 AM, Sage Weil wrote:
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> On Wed, 26 Sep 2012, hemant surale wrote:
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> Hi Dan ,
>>>>>>>>> I have set replication factor to 3 of pool 'newbyh' . Then
>>>>>>>>> when
>>>>>>>>> i
>>>>>>>>> tried to execute cmds told by you I got that it reported like
>>>>>>>>>
>>>>>>>>> ----------------------------------------------------
>>>>>>>>> root@third-virtual-machine:~# osdmaptool --test-map-object Obj1
>>>>>>>>> osdmap
>>>>>>>>> osdmaptool: osdmap file 'osdmap'
>>>>>>>>> object 'Obj1' -> 0.c3c4 -> [0,1]
>>>>>>>>> -----------------------------------------------------
>>>>>>>>>
>>>>>>>>> I even checked manually the dirs at every node . it shows proper data
>>>>>>>>> available within osd0,osd1 & osd2 . ( i hve 3 node cluster using VM
>>>>>>>>> Ceph v0.36 )
>>>>>>>>>
>>>>>>>>> So my questions is why in above execution it shows that Obj1 is at
>>>>>>>>> [0,1] .. it should report all nodes like [0,1,2] .
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> The --test-map-object is currently somewhat useless because it assumes
>>>>>>>> pool 0 ('data'), and your object is probably in a different pool.
>>>>>>>>
>>>>>>>> sage
>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> -
>>>>>>>>> Hemant Surale.
>>>>>>>>>
>>>>>>>>> On Wed, Sep 26, 2012 at 2:04 AM, Dan Mick <dan.mick@xxxxxxxxxxx>
>>>>>>>>> wrote:
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> Hemant:
>>>>>>>>>>
>>>>>>>>>> Yes, you can. Use ceph osd getmap -o <file> to get the OSD map, and
>>>>>>>>>> then
>>>>>>>>>> use osdmaptool --find-object-map <objectname> <file> to output the
>>>>>>>>>> PG the object hashes to and the list of OSDs that PG maps to
>>>>>>>>>> (primary
>>>>>>>>>> first):
>>>>>>>>>>
>>>>>>>>>> $ ceph osd getmap -o osdmap
>>>>>>>>>> got osdmap epoch 59
>>>>>>>>>> $ osdmaptool --test-map-object dmick.rbd osdmap
>>>>>>>>>> osdmaptool: osdmap file 'osdmap'
>>>>>>>>>> object 'dmick.rbd' -> 0.69c8 -> [3,1]
>>>>>>>>>>
>>>>>>>>>> shows dmick.rbd mapping to pg 0.69c8, which in turn maps to OSDs 3
>>>>>>>>>> and
>>>>>>>>>> 1, 3
>>>>>>>>>> being the primary.
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> On 09/25/2012 02:30 AM, hemant surale wrote:
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> Hi Community,
>>>>>>>>>>> Is it possible to identify where exactly primary copy
>>>>>>>>>>> of
>>>>>>>>>>> obj
>>>>>>>>>>> is stored ? I am using crushmaps to use specific osds for data
>>>>>>>>>>> placement but i want to knw the primary capoy location. Or I need
>>>>>>>>>>> to
>>>>>>>>>>> replace pseudo random function by some deterministic function to
>>>>>>>>>>> guide
>>>>>>>>>>> ceph to utilize specific osd?
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> Regards,
>>>>>>>>>>> Hemant Surale.
>>>>>>>>>>> --
>>>>>>>>>>> To unsubscribe from this list: send the line "unsubscribe
>>>>>>>>>>> ceph-devel"
>>>>>>>>>>> in
>>>>>>>>>>> the body of a message to majordomo@xxxxxxxxxxxxxxx
>>>>>>>>>>> More majordomo info at http://vger.kernel.org/majordomo-info.html
>>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>> --
>>>>>>>>> To unsubscribe from this list: send the line "unsubscribe ceph-devel"
>>>>>>>>> in
>>>>>>>>> the body of a message to majordomo@xxxxxxxxxxxxxxx
>>>>>>>>> More majordomo info at http://vger.kernel.org/majordomo-info.html
>>>>>>>>>
>>>>>>>>>
>>>>>>>
>>>>>
>>>
> --
> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
> the body of a message to majordomo@xxxxxxxxxxxxxxx
> More majordomo info at http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
