Re: [ceph-users] Object location

[Dan Mick]

Please don't remove the list. 

No, you can't choose a particular OSD for a particular object access. Why do 
you believe this is important?

On Mar 31, 2013, at 5:34 AM, Waed Bataineh  wrote:

> oh thank you Dan,
> I think it's more clear right now for me. but still can i force it to save 
> the file as object in specific osd? 
> 
> Best regards. 
> 
> 
> On Sat, Mar 30, 2013 at 10:17 AM, Dan Mick  wrote:
>> Keep in mind also that the osd map command doesn't require the object to 
>> exist. So you can run some maps and see which objects will land where if and 
>> when they are created. 
>> 
>> On Mar 29, 2013, at 9:32 PM, Waed Bataineh  wrote:
>> 
>>> Dan,
>>> 
>>> well exactly if i used the mapping command line i have confusion about some 
>>> issues if i may :
>>> _ exactly what it meant by obj in  rados -p  put   , is 
>>> this obj consider like the first part of the file and pointing  to the 
>>> other object?
>>> _how i can get the file back ?
>>> _Finally, if i have two OSDs n no replica when do the mapping it always 
>>> store at osd.0, is there a way to mak it store on osd.1 or it will continue 
>>> till osd.0 will reach nearfull ? 
>>> 
>>> Thank you.
>>> 
>>> 
>>> On Sat, Mar 30, 2013 at 7:15 AM, Dan Mick  wrote:
 so "rados -p  put  " is one way to create an object. 
 There are also C/C++/Python librados bindings for writing programs to do 
 so.
 
 
 On 03/28/2013 05:30 AM, Waed Bataineh wrote:
> Object Store.
> 
> 
> On Thu, Mar 28, 2013 at 11:51 AM, Sebastien Han
> mailto:sebastien@enovance.com>> wrote:
> 
> Hi,
> 
> It depends, what do you to use from Ceph? Object store? Block
> device? Distributed filesystem?
> 
> 
> Sébastien Han
> Cloud Engineer
> 
> "Always give 100%. Unless you're giving blood."
> 
> 
> 
> 
> 
> 
> 
> 
> 
> PHONE : +33 (0)1 49 70 99 72 – MOBILE : +33 (0)6 52 84 44 70
> EMAIL : sebastien@enovance.com
>  – SKYPE : han.sbastien
> 
> ADDRESS : 10, rue de la Victoire – 75009 Paris
> WEB : www.enovance.com  – TWITTER : @enovance
> 
> 
> On Mar 28, 2013, at 9:31 AM, Waed Bataineh  > wrote:
> 
>> Thank you both.
>> 
>> well , no more error message :)
>> 
>> If i may, i still have some confusion about how to start operating
>> on ceph!
>> 
>> _how can i start storing files that on the client at the server?
>> correct me if i'm wrong, at client side i'll save the files i need
>> to store at the mount point, then the server should deal with this
>> files and start ceph work!
>> 
>> _in case that what is really going, how i can get to the path the
>> files have been stored in? this is actually my major problem, dont
>> know the path that lead me to the files or objects!
>> 
>> Hope I'm making my self clear!
>> 
>> Thank you again.
>> 
>> 
>> On Thu, Mar 28, 2013 at 3:20 AM, John
>> Wilkins > > wrote:
>> Already done. Missed it at first. Thanks! :)
>> 
>> 
>> On Wed, Mar 27, 2013 at 5:17 PM, Sebastien
>> Han > > wrote:
>> Hi John,
>> 
>> Could you also please update the generic usage at the beginning of
>> the section?
>> 
>> Thanks!
>> 
>> ps: I tried to push a change, but something went wrong with Github...
>> 
>> 
>> Sébastien Han
>> Cloud Engineer
>> 
>> "Always give 100%. Unless you're giving blood."
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> PHONE : +33 (0)1 49 70 99 72 – MOBILE : +33 (0)6 52 84 44 70
>> EMAIL : sebastien@enovance.com
>>  – SKYPE : han.sbastien
>> 
>> ADDRESS : 10, rue de la Victoire – 75009 Paris
>> WEB : www.enovance.com  – TWITTER : 
>> @enovance
>> 
>> 
>> On Mar 28, 2013, at 12:52 AM, John Wilkins
>> mailto:john.wilk...@inktank.com>> wrote:
>> 
>>> Sebastian,
>>> 
>>> You're correct. The usage has changed, so I've updated the doc to
>>> reflect it. I ran through the procedure and it worked fine for
>>> me. Sorry Waed. I hope that's all that was wrong.
>>> 
>>> On Wed, Mar 27, 2013 at 3:48 PM, Sebastien Han
>>> mailto:sebastien@enovance.com>>
>>> 
>>> wrote:
>>> Arf sorry, not 'odd' but 'osd' of course. (thanks autocompletion…)
>>> 
>>> 
>>> Sébastien Han
>>> Cloud Engineer
>>> 
>>> "Alwa

Re: [ceph-users] Object location

[Waed Bataineh]

sorry my mistake.

well i'm still learning about Ceph here but consider in the environment of
two osds and i'm using the object storage and no replica used, then in some
matter i needed to clean up one osd and move the content to the other osd?
cant do it ?

Thank you.


On Sun, Mar 31, 2013 at 10:08 PM, Dan Mick  wrote:

> Please don't remove the list.
>
> No, you can't choose a particular OSD for a particular object access. Why
> do you believe this is important?
>
> On Mar 31, 2013, at 5:34 AM, Waed Bataineh 
> wrote:
>
> oh thank you Dan,
> I think it's more clear right now for me. but still can i force it to save
> the file as object in specific osd?
>
> Best regards.
>
>
> On Sat, Mar 30, 2013 at 10:17 AM, Dan Mick  wrote:
>
>> Keep in mind also that the osd map command doesn't require the object to
>> exist. So you can run some maps and see which objects will land where if
>> and when they are created.
>>
>> On Mar 29, 2013, at 9:32 PM, Waed Bataineh 
>> wrote:
>>
>> Dan,
>>
>> well exactly if i used the mapping command line i have confusion about
>> some issues if i may :
>> _ exactly what it meant by obj in * rados -p  put  * ,
>> is this obj consider like the first part of the file and pointing  to the
>> other object?
>> _how i can get the file back ?
>> _Finally, if i have two OSDs n no replica when do the mapping it always
>> store at osd.0, is there a way to mak it store on osd.1 or it will continue
>> till osd.0 will reach nearfull ?
>>
>> Thank you.
>>
>>
>> On Sat, Mar 30, 2013 at 7:15 AM, Dan Mick  wrote:
>>
>>> so "rados -p  put  " is one way to create an object.
>>> There are also C/C++/Python librados bindings for writing programs to do so.
>>>
>>>
>>> On 03/28/2013 05:30 AM, Waed Bataineh wrote:
>>>
 Object Store.


 On Thu, Mar 28, 2013 at 11:51 AM, Sebastien Han
 >>> >
 wrote:

 Hi,

 It depends, what do you to use from Ceph? Object store? Block
 device? Distributed filesystem?

 
 Sébastien Han
 Cloud Engineer

 "Always give 100%. Unless you're giving blood."









 PHONE : +33 (0)1 49 70 99 72 – MOBILE : +33 (0)6 52 84 44 70
 EMAIL : sebastien@enovance.com
 
 – SKYPE : han.sbastien

 ADDRESS : 10, rue de la Victoire – 75009 Paris
 WEB : www.enovance.com  – TWITTER :
 @enovance


 On Mar 28, 2013, at 9:31 AM, Waed Bataineh >>> >
 wrote:

  Thank you both.
>
> well , no more error message :)
>
> If i may, i still have some confusion about how to start operating
> on ceph!
>
> _how can i start storing files that on the client at the server?
> correct me if i'm wrong, at client side i'll save the files i need
> to store at the mount point, then the server should deal with this
> files and start ceph work!
>
> _in case that what is really going, how i can get to the path the
> files have been stored in? this is actually my major problem, dont
> know the path that lead me to the files or objects!
>
> Hope I'm making my self clear!
>
> Thank you again.
>
>
> On Thu, Mar 28, 2013 at 3:20 AM, John
> Wilkins  >
> wrote:
> Already done. Missed it at first. Thanks! :)
>
>
> On Wed, Mar 27, 2013 at 5:17 PM, Sebastien
> Han  >
> wrote:
> Hi John,
>
> Could you also please update the generic usage at the beginning of
> the section?
>
> Thanks!
>
> ps: I tried to push a change, but something went wrong with
> Github...
>
> 
> Sébastien Han
> Cloud Engineer
>
> "Always give 100%. Unless you're giving blood."
>
>
>
>
>
>
>
>
>
> PHONE : +33 (0)1 49 70 99 72 – MOBILE : +33 (0)6 52 84 44 70
> EMAIL : sebastien@enovance.com
> 
> – SKYPE : han.sbastien
>
> ADDRESS : 10, rue de la Victoire – 75009 Paris
> WEB : www.enovance.com  – TWITTER :
> @enovance
>
>
> On Mar 28, 2013, at 12:52 AM, John Wilkins
>  >
> wrote:
>
>  Sebastian,
>>
>> You're correct. The usage has changed, so I've updated the doc to
>> reflect it. I ran through the procedure and it worked fine for
>> me. Sorry Waed. I hope that's all that was wrong.
>>
>> On Wed, Mar 27, 2013 at 3:48 PM, Sebastien Han
>> > 

[ceph-users] under performing osd, where to look ?

[Matthieu Patou]

Hi,

I was doing some testing with iozone and found that performance of an 
exported rdb volume where 1/3 of the performance of the hard drives.

I was expecting to have a performance penalty but not so important.
I suspect something is not correct in the configuration but I can't what 
exactly.


My test setup consists of 2 OSD with one drive each they have 2 network 
(public/private).
On osd0 and osd1 I'm using a BCM5751 gigabit card for the public network 
and a intel 82541PI for the private.

On the client I'm using a RTL8111/8168B for the public network.

Osd0 & Osd1 are connect directly with a cable for the private network 
and they are connected through one switch for the public network.


Where could the bottleneck be located ?

Matthieu.


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