[Gluster-users] GeoReplication takes too log in hybrid crawl and no sync happens in changelog mode for 2x3 volume.

Sunny Kumar sunkumar at redhat.com
Wed Feb 12 17:37:53 UTC 2020

Hi Deepu,

I mean to say rename and deletes on master node will not be sync to
slave node during hybrid crawl.


On Wed, Feb 12, 2020 at 5:24 PM deepu srinivasan <sdeepugd at gmail.com> wrote:
> Thank you for the clarification.  If when will the rename and delete will happen to the slave end.
> On Wed, Feb 12, 2020, 10:49 PM Sunny Kumar <sunkumar at redhat.com> wrote:
>> Hi Deepu,
>> Basically in Hybrid crawl, the sync daemon will crawl through full
>> file system in your case 127GB to generating pseudo changelog  which
>> it uses to sync the data, that's why it can take some time.
>> On Wed, Feb 12, 2020 at 8:15 AM deepu srinivasan <sdeepugd at gmail.com> wrote:
>> >
>> > Hi Users/Developers
>> > We have a geo-replication session and it takes too long in hybrid crawl mode. With 127GB of data in our volume, it takes nearly two days to change to changelog mode. And during changelog, the sync does not happen properly. How to debug the issue and does it take this long to hybrid mode.
>> What do you mean by sync not happening properly.
>> Also please note when the sync is in hybrid mode it will not sync
>> renames and detetes to  the slave volume.
>> > We have a 2x3 volume and we have enabled geo replication in it .
>> /sunny

More information about the Gluster-users mailing list