Additional info: For example: - Open a port for ssh / rsync - Configure the nodes for passwordless SSH access for the replication. - (from Humble) “ideally ssh port ( 2222) will be opened from container. We should have a mention about the port in the doc which we have opened. We can think about making it a configurable port by an env var, but need to revisit the possibilities from cns-deploy utility pov.”
Prasanth, According to the eng, this feature is test-only effort and the writer needs to co-ordinate with the QA for information. Could you please let me know the information to be added on how to set up CNS 3.5 volumes for geo-replication to a non-CNS remote site? If you are not the QA contact for this feature, could you move the need_info to the concerned QA engineer? Thanks!
The only mention what we have to do here is, the ssh port which have opened in the gluster container is '2222'. Other than that we can just make use of our geo replication content as such. Apeksha, being you are the one who qualified this feature for CNS 3.5, do you want to add any extra details here ?
We also need to open ports in range this -{49152..49251} using firewall cmd, to setup geo-rep session. Rest we can follow the RhGS doc to setup geo-rep, it all remains the same.
(In reply to Apeksha from comment #7) > We also need to open ports in range this -{49152..49251} using firewall cmd, > to setup geo-rep session. Rest we can follow the RhGS doc to setup geo-rep, > it all remains the same. Why are you opening just 100 ports here, iic, we atleast open 500 ports in gluster containers? Also how did you test geo replication ? Was the slave volume stayed in openshift cluster or outside openshift cluster?
I was getting an error with respect to port 49152, so Raghvendra Talur suggested to open atleast 100 ports. These ports were opened on the RHGS nodes outside the openshift cluster. Geo-rep setup: The master volume on the openshift cluster, whereas the slave volume stayed outside the openshift cluster
(In reply to Apeksha from comment #9) > I was getting an error with respect to port 49152, so Raghvendra Talur > suggested to open atleast 100 ports. These ports were opened on the RHGS > nodes outside the openshift cluster. Well, we always say open atleast 500 ports: Check iptables command in the doc https://access.redhat.com/documentation/en-us/red_hat_gluster_storage/3.1/html-single/container-native_storage_for_openshift_container_platform_3.4/ > > Geo-rep setup: > The master volume on the openshift cluster, whereas the slave volume stayed > outside the openshift cluster Thanks for the clartifiaction. So, Divya can we carve out a guide based on this ?
Thanks for the inputs Humble and Apeksha. I will update the guide based on your inputs and reach out to you if required.
(In reply to Apeksha from comment #7) > We also need to open ports in range this -{49152..49251} using firewall cmd, > to setup geo-rep session. Rest we can follow the RhGS doc to setup geo-rep, > it all remains the same. Humble, Apeksha, According to our existing documentation, in the "3.2.3. Red Hat OpenShift Container Platform Requirements" section [1], we have already asked users to open ports 2222 and 49152-49664. So, I don't think we need to add any special instructions for geo-replication. Could you please confirm if my understanding is correct? If yes, do we close this bug? [1] http://ccs-jenkins.gsslab.brq.redhat.com:8080/job/doc-Red_Hat_Gluster_Storage-3.5-Container_Native_Storage_with_OpenShift_Platform-branch-master/lastSuccessfulBuild/artifact/tmp/en-US/html-single/index.html#idm140643980091264
Added information about geo-replication based on Comment 15 and 16. Link to the doc: http://ccs-jenkins.gsslab.brq.redhat.com:8080/job/doc-Red_Hat_Gluster_Storage-3.5-Container_Native_Storage_with_OpenShift_Platform-branch-master/lastSuccessfulBuild/artifact/tmp/en-US/html-single/index.html#chap-Documentation-Red_Hat_Gluster_Storage_Container_Native_with_OpenShift_Platform-Gluster_Pod_Operations