[Drbd-dev] Question About current uuid in drbd9

[Kim, Sung-eun]

Hi, Philipp

Thank you for your answer.

But I don't understand.

If a primary's lowest bit must be 1 as your comment,
*0006A4E400035276(primary's
current uuid)* is not (6 is 011*0*), lowest bit is zero.

primary's uuid and secondary's uuid is same (0006A4E400035276, is even
number.) in my case.

I think that 0006A4E40003527*7* is a correct value. Isn't it?

This test result is reproduced in drbd9, and a different part with drbd8.x.

in drbd8.x, primary's current uuid is a always odd number(lowest bit is 1).



scenario:

after initial full sync, change node1's role to secondary, and again change
to primary.

at this point, node1(pri)'s current uuid is equal to secondary from
'drbdadm get-gi'.




Thanks.

Best Regards,
Sung-eun Kim







[image: 설명: logo]

Technical Research Center / DRBD Team

Principal Research Engineer

*Sung-eun Kim               sekim at mantech.co.kr <sekim at mantech.co.kr>*

202 Fl, Seoulforest Kolon digital Tower, 308-4, Seongsudong 2ga,
Seongdong-gu, Seoul, Korea

Tel : 02-2136-6913 / Fax : 02-575-4858 / Call Center : 1833-7790

http://www.mantech.co.kr

[image: 본문 이미지 1]

2016-08-25 1:59 GMT+09:00 Philipp Reisner <philipp.reisner at linbit.com>:

> Hi Sung-eun,
>
> Only 63 bits are random. We encode the current role (Primary = 1)
> into the lowest bit. When comparing the UUIDs the lowest bit
> is ignored.
>
> From top of my head, one of the auto-recovery-after-split-brain
> policies looks at this bit.
>
> This is normal behavior and expected.
>
> best regards,
>  phil
>
> Am Mittwoch, 24. August 2016, 16:41:05 CEST schrieb Kim, Sung-eun:
> > Hello,
> >
> >
> > I have a question about current uuid in drbd9
> >
> > After initial full-sync, primary's current uuid is one more than
> > secondary's current uuid.
> >
> > ex) from drbdadm get-gi r0
> >
> > primary *0005F0D20002F86D*
> >
> > :0000000000000000:0000000000000000:0000000000000000:1:1:1:0:0:0:1:0:0:0
> >
> > secondary
> > *0005F0D20002F86C*
> >
> > :0000000000000000:0000000000000000:0000000000000000:1:1:0:0:0:0:1:0:0:0
> >
> > But I changed primary's role to secondary , and again changed to primary.
> > and replication volume disk I/O is processed.
> >
> > At this point, primary's current uuid is equal to secondary's current
> > uuid.(not one more than secondary's)
> >
> > primary
> > *0006A4E400035276*
> >
> > :0000000000000000:0000000000000000:0000000000000000:1:1:1:0:0:0:1:0:0:0
> >
> > secondary
> > *0006A4E400035276*
> >
> > :0000000000000000:0000000000000000:0000000000000000:1:1:0:0:0:0:1:0:0:0
> >
> > Is this Bug? , or normal situation?(intended implementation in drbd9?).
> >
> > Could you let me about this result?
> >
> > Thanks.
> >
> >
> > Best Regards,
> > Sung-eun Kim
> >
> > [image: 설명: logo]
> >
> > Technical Research Center / DRBD Team
> >
> > Principal Research Engineer
> >
> > *Sung-eun Kim               sekim at mantech.co.kr <sekim at mantech.co.kr>*
> >
> > 202 Fl, Seoulforest Kolon digital Tower, 308-4, Seongsudong 2ga,
> > Seongdong-gu, Seoul, Korea
> >
> > Tel : 02-2136-6913 / Fax : 02-575-4858 / Call Center : 1833-7790
> >
> > http://www.mantech.co.kr
> >
> > [image: 본문 이미지 1]
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linbit.com/pipermail/drbd-dev/attachments/20160825/3009b09d/attachment-0001.htm>