[Starlingx-discuss] Second controller on degraded for long time - Duplex Virtual

Hu, Yong yong.hu at intel.com
Fri Nov 16 01:35:54 UTC 2018


Hi José,
I checked the alarm you pasted and found it seemed normal that the data sync b/w 2 controllers was in progress - "data-syncing":

Service group controller-services degraded; drbd-extension(enabled-standby, degraded, data-syncing),    
|          | drbd-cinder(enabled-standby, degraded, data-syncing)

*If you still have such an environment, you can check host status once again after such a long time (hours/days).*

Moreover, 2 other comments:
1). you can use "drbd-overview" to check the sync percentage, whenever you see "degraded" state caused triggered by "data-syncing" alarm.
As Stefan pointed out the sync might take very long time if the partition size is big and network speed is relatively slow.
2). The sync might fail eventually in the case that PV sizes on controller-0 and controller-1 is not the same (or roughly identically sized). 
In such a failure case, you can also see the details by "drbd-overview".
   

Regards,
Yong

On 13/11/2018, 12:56 PM, "Hu, Yong" <yong.hu at intel.com> wrote:

    Hi José,
    I ever saw this kind of issue when I added a very big disk partition (like 900 GB) on controller node by "host-lvg-modify".
    
    And you might see some info meaningful by "fm alarm-list". In my case the message was about inconsistent in drbd-cinder.
    
    Regards,
    Yong  
    
    On 12/11/2018, 9:29 PM, "Perez Carranza, Jose" <jose.perez.carranza at intel.com> wrote:
    
        Hi Cindy / Austin 
        
        As I mentioned on my previous mail, I did not see the behavior mentioned on the bug, in my case the controller-1 becomes "Available" after the synchronization is finished and did not goes back again to "Degraded" mode.  All my tests were done on an virtual environment. 
        
        Regards,
        José
        
        
        > -----Original Message-----
        > From: Sun, Austin
        > Sent: Friday, November 9, 2018 7:10 PM
        > To: Xie, Cindy <cindy.xie at intel.com>; Perez Carranza, Jose
        > <jose.perez.carranza at intel.com>; Dinescu, Stefan
        > <Stefan.Dinescu at windriver.com>; 'starlingx-discuss at lists.starlingx.io'
        > <starlingx-discuss at lists.starlingx.io>
        > Subject: RE: [Starlingx-discuss] Second controller on degraded for long time -
        > Duplex Virtual
        > 
        > Hi Cindy:
        >    Mhg raised bug#1798836 and I worked with him.  I cannot reproduce this
        > issue.  bug#1798836 is using static IPv4 , not dynamic .
        >    Mhg's  setup has been changed to do other test.
        > @Jose:    could  you raise bug and attach logs ( /var/log/*)  in controller-
        > 0/controller-1 ?
        > 
        > Thanks.
        > BR
        > Austin Sun.
        > 
        > 
        > 
        > -----Original Message-----
        > From: Xie, Cindy
        > Sent: Saturday, November 10, 2018 8:12 AM
        > To: Perez Carranza, Jose <jose.perez.carranza at intel.com>; Sun, Austin
        > <austin.sun at intel.com>; Dinescu, Stefan <Stefan.Dinescu at windriver.com>;
        > 'starlingx-discuss at lists.starlingx.io' <starlingx-discuss at lists.starlingx.io>
        > Subject: RE: [Starlingx-discuss] Second controller on degraded for long time -
        > Duplex Virtual
        > 
        > So Jose, you think the issue you encountered is the same issue as
        > bug#1798836, my understanding correct?
        > 
        > @Austin who is working on this bug may contact you for more details
        > including log files.
        > 
        > Thx. - cindy
        > 
        > -----Original Message-----
        > From: Perez Carranza, Jose
        > Sent: Saturday, November 10, 2018 1:37 AM
        > To: Xie, Cindy <cindy.xie at intel.com>; Dinescu, Stefan
        > <Stefan.Dinescu at windriver.com>; 'starlingx-discuss at lists.starlingx.io'
        > <starlingx-discuss at lists.starlingx.io>
        > Subject: RE: [Starlingx-discuss] Second controller on degraded for long time -
        > Duplex Virtual
        > 
        > Hi Cindy
        > 
        > Actually I did not  fill a Launchpad, I just sent the question to the mailing list,
        > in my case controller-1 gets available after sync (~1 hour) but I did not
        > noticed that after that a controller goes back to degraded mode again.
        > 
        > Regards,
        > José
        > 
        > 
        > > -----Original Message-----
        > > From: Xie, Cindy
        > > Sent: Thursday, November 8, 2018 5:16 PM
        > > To: Dinescu, Stefan <Stefan.Dinescu at windriver.com>; Perez Carranza,
        > > Jose <jose.perez.carranza at intel.com>; 'starlingx-discuss at lists.starlingx.io'
        > > <starlingx-discuss at lists.starlingx.io>
        > > Subject: RE: [Starlingx-discuss] Second controller on degraded for
        > > long time - Duplex Virtual
        > >
        > > Jose,
        > > Can you check if you reported issue is similar as the existing one
        > > reported
        > > here:
        > >
        > > https://bugs.launchpad.net/starlingx/+bug/1798836
        > >
        > > thx. - cindy
        > >
        > > -----Original Message-----
        > > From: Dinescu, Stefan [mailto:Stefan.Dinescu at windriver.com]
        > > Sent: Thursday, November 8, 2018 11:01 PM
        > > To: Perez Carranza, Jose <jose.perez.carranza at intel.com>; 'starlingx-
        > > discuss at lists.starlingx.io' <starlingx-discuss at lists.starlingx.io>
        > > Subject: Re: [Starlingx-discuss] Second controller on degraded for
        > > long time - Duplex Virtual
        > >
        > > Hello,
        > >
        > > Depending on the size of your cinder partition and network speed
        > > between the two controller nodes, cinder make take some time to sync
        > > as it is usually the biggest partition.
        > >
        > > you can use "drbd-overview" to see the progress of the sync.
        > >
        > > Thanks,
        > > Stefan
        > > ________________________________________
        > > From: Perez Carranza, Jose [jose.perez.carranza at intel.com]
        > > Sent: Thursday, November 08, 2018 4:56 PM
        > > To: 'starlingx-discuss at lists.starlingx.io'
        > > Subject: [Starlingx-discuss] Second controller on degraded for long
        > > time - Duplex Virtual
        > >
        > > Hi
        > >
        > > I setup a Duplex virtual environment and I noticed that after
        > > installation and provisioning of second controller (controller-1) this
        > > one stays in Degraded for long time (~ 1 hour), and alarm 400.001 is
        > > being displayed [1], Is this normal behavior? Is here a way to take
        > > out the controller-1 from 'degraded' to 'available' faster?
        > >
        > > 1- http://paste.openstack.org/show/734425/
        > >
        > >
        > > Regards,
        > > José
        > >
        > >
        > >
        > > _______________________________________________
        > > Starlingx-discuss mailing list
        > > Starlingx-discuss at lists.starlingx.io
        > > http://lists.starlingx.io/cgi-bin/mailman/listinfo/starlingx-discuss
        > >
        > > _______________________________________________
        > > Starlingx-discuss mailing list
        > > Starlingx-discuss at lists.starlingx.io
        > > http://lists.starlingx.io/cgi-bin/mailman/listinfo/starlingx-discuss
        _______________________________________________
        Starlingx-discuss mailing list
        Starlingx-discuss at lists.starlingx.io
        http://lists.starlingx.io/cgi-bin/mailman/listinfo/starlingx-discuss
        
    
    _______________________________________________
    Starlingx-discuss mailing list
    Starlingx-discuss at lists.starlingx.io
    http://lists.starlingx.io/cgi-bin/mailman/listinfo/starlingx-discuss
    



More information about the Starlingx-discuss mailing list