Re: how ceph mon works

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Tue, 26 Apr 2016 23:31:39 +0200 (CEST) Wido den Hollander wrote:

> 
> > Op 26 april 2016 om 23:24 schreef yang sheng <forsaks.30@xxxxxxxxx>:
> > 
> > 
> > Hi
> > 
> > according to ceph docs, it recommend 3 monitors as least. All the
> > clients will contact monitors first to get the ceph map and connect
> > the osd.
> > 
> 
> Yes, indeed.
> 
> > I am curious that if I have 3 monitors, are these monitors run in
> > master-master mode or master-slave mode?  In another word, will clients
> > talk to any of these 3 monitors or all clients can only talk to one
> > monitor and if that monitor is unreachable, than they talk to the
> > second monitor?
> > 
> 
> During the eleOn 04/26/2016 04:59 PM, M.Z.Medina wrote:
> anti mosquito ction process one monitor will become the 'leader' and
> others will become 'peons'.
> 
This is also by default the monitor with the lowest IP address, so until
one day that may become configurable, keep this in mind and put dedicated
MONs with faster HW at the low IP address(es).

> A client can talk to each monitor, but the peons will forward commands
> to the leader which makes the decisions inside the cluster.
> 
AFAIK this only goes for changes (i.e. adding an OSD, changing weights),
getting information will be handled by the MON in question with no need
for it to contact the leader.

> Clients are aware of all the Monitors in a cluster, so if a monitor goes
> down they can contact the other one.
> 
While true, clients basically contact MONs in a round-robin fashion
(randomized or not, I don't know). 

While this works, it of course introduces delays when they try to reach a
dead MON.
I've always felt that something more elegant would be nice here, as in not
sending any real requests to a MON that hasn't been responsive and retry
that MON with a dummy request every minute or so and if successful add it
back to the to "good" pool.

Regards,

Christian
> Wido
> 
> > 
> > thanks
> > _______________________________________________
> > 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
> 


-- 
Christian Balzer        Network/Systems Engineer                
chibi@xxxxxxx   	Global OnLine Japan/Rakuten Communications
http://www.gol.com/
_______________________________________________
ceph-users mailing list
ceph-users@xxxxxxxxxxxxxx
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com



[Index of Archives]     [Information on CEPH]     [Linux Filesystem Development]     [Ceph Development]     [Ceph Large]     [Ceph Dev]     [Linux USB Development]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]     [xfs]


  Powered by Linux