Hide Forgot
Document URL: https://access.redhat.com/documentation/en/red-hat-gluster-storage/3.1/paged/deployment-guide-for-containerized-red-hat-gluster-storage-in-openshift-enterprise/ I have multiple comments after reviewing the complete document. Here are my comments and feedback. You may have to work with development team to get inputs on few comments. comment # 1 section 3.2.1.1, When deploying this solution on Atomic OpenShift it is necessary to deploy a client outside the cluster to install the heketi RPMs. It is optional, though an accepted best practice, to use the OpenShift Master as a client to execute the 'oc' commands across the cluster when installing OpenShift. Generally, this is setup as a non-scheduled node in the cluster. This is the default configuration when using the OpenShift installer. A user can also choose to install their client on their local machine to access the cluster remotely. For more information, see https://access.redhat.com/documentation/en/openshift-enterprise/version-3.2/cli-reference/#installing-the-cli. When section 3.2.1 is about configuring openshift on RHEL, there is no need for any mention of rhel atomic. anything related to atomic should be moved to 3.2.2 e.g., section 3.2.1.1 starts with atomic host! Generally, this is setup as a non-scheduled node in the cluster. --> this to be replaced by openshift master node Overall, it would be better to describe about configuring a client and then finally mentioning about master itself acting as client in case of openshift on rhel comment # 2 architecture diagram doesn't really represent RHGS converged with openshift. A node should have gluster and applicattion pods residing together and multiple such nodes should be shown to form a cluster. current diagram doesn't depict the converged solution correctly. comment # 3: section 3.2.1.2 --> Information about 'subscribe to openshift repo' is provided when a client is configured separately, same will have to be done for master as well. the way the steps unfold is misleading. comment # 4: Either a note or link to the setup of openshift cluster has to be provided before section 3.2.3 comment # 5: After, Execute the following command to reload the iptables: # systemctl reload iptables A validation step can be added, iptables -L on all nodes to validate if actually iptables have been updated. comment # 6: section 3.2.3, last point, Ensure the nodes have the valid ports opened for communicating with Red Hat Gluster Storage. See the iptables configuration task in Step 1 of Section 4.1, “Preparing the Red Hat OpenShift Container Platform Cluster” for more information. The link doesn't provide any information on ports that are opened. comment # 7: section 4.1, note under step 3 says, If a router already exists, proceed to Step 5. It would be good if appropriate command is provided to see if router is already configured. comment # 8: section 4.1, under step 3 --> while step 1&2 has to be run on client, step 3&4 has to be run on master node. comment # 9: section 4.1, under step 3 --> step 3 & 4 -> what do we do if we have more than 1 master node? Information on whether to edit on all master nodes need to be updated comment # 10: section 4.1, step 3 has further 4 sub-steps. they are also numbered 1,2,3,4. Instead (i), (ii), (iii) and (iv) can be used. This change has to be done all over the document. comment # 11: section 4.1, before step 4, give the command to check if router is running. If this is done, note under step 4.1 is not required. comment # 12: section 4.1, note under step 5, Red Hat Gluster Storage container can only run in privileged mode. There is nothing one can do with this information. do we have to check if the pods are in previleged mode? if they are not in previlieged mode, what needs to be done? comment # 13: step 4.2, Title needs to be more informative, what template are we creating? A brief description is neeed on what are these templates and why are we creating these templates comment # 14: step 4.3, Title needs to be more informative. comment # 15: note under section 4.4, For simplicity, the file only sets up 3 nodes with 4 drives each. why do we have this line? We neither have the sample screenshot nor a 'file' there. comment # 16: A brief description on heketi before getting into the commands to deploy heketi would be great comment # 17: step 9, only after creating pvc.json file the command can be executed. Hence provide a step to create the file first.
The comments mentioned in the description are divided amongst Divya and me. I have worked on Comments 11 to 17 and Divya on the rest. The following update is for comments 11 to 17: http://ccs-jenkins.gsslab.brq.redhat.com:8080/job/doc-Red_Hat_Gluster_Storage-3.4-Container_Native_Storage_with_OpenShift_Platform-branch-bz-1380076-part2-Comments11-17/lastSuccessfulBuild/artifact/tmp/en-US/html-single/index.html comment # 11: section 4.1, before step 4, give the command to check if router is running. If this is done, note under step 4.1 is not required. ---> Added step 4 to verify if the router is set up. The movement of the note, will be taken care of by Divya for one of the comments she is working on. comment # 12: section 4.1, note under step 5, Red Hat Gluster Storage container can only run in privileged mode. There is nothing one can do with this information. do we have to check if the pods are in previleged mode? if they are not in previlieged mode, what needs to be done? ---> Added note as part of the step. comment # 13: step 4.2, Title needs to be more informative, what template are we creating? A brief description is neeed on what are these templates and why are we creating these templates ---> As discussed in the meeting with Karthik, it was decided that no changes are required comment # 14: step 4.3, Title needs to be more informative. ---> As discussed in the meeting with Karthik, it was decided that no changes are required comment # 15: note under section 4.4, For simplicity, the file only sets up 3 nodes with 4 drives each. why do we have this line? We neither have the sample screenshot nor a 'file' there. ---> The note is now added as a para just after the example file. The file name is explicitly called out to avoid any confusion comment # 16: A brief description on heketi before getting into the commands to deploy heketi would be great ---> As discussed in the meeting with Karthik, it was decided that no changes are required, as a brief description is already provided at the beginning of the section and also in chapter 2 comment # 17: step 9, only after creating pvc.json file the command can be executed. Hence provide a step to create the file first. ---> In Section 5.2 Static Provisioning of Volumes, step 9 is added to create a persistent volume claim and step 10 is about binding the persistent volume to the pvc
Thanks Karthick and Ashiq for all the help! Link to the updated doc: http://ccs-jenkins.gsslab.brq.redhat.com:8080/job/doc-Red_Hat_Gluster_Storage-3.4-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-Support_Requirements_2 The following update is for comments 1 to 10: comment # 1 section 3.2.1.1, When deploying this solution on Atomic OpenShift it is necessary to deploy a client outside the cluster to install the heketi RPMs. It is optional, though an accepted best practice, to use the OpenShift Master as a client to execute the 'oc' commands across the cluster when installing OpenShift. Generally, this is setup as a non-scheduled node in the cluster. This is the default configuration when using the OpenShift installer. A user can also choose to install their client on their local machine to access the cluster remotely. For more information, see https://access.redhat.com/documentation/en/openshift-enterprise/version-3.2/cli-reference/#installing-the-cli. When section 3.2.1 is about configuring openshift on RHEL, there is no need for any mention of rhel atomic. anything related to atomic should be moved to 3.2.2 e.g., section 3.2.1.1 starts with atomic host! Generally, this is setup as a non-scheduled node in the cluster. --> this to be replaced by openshift master node Overall, it would be better to describe about configuring a client and then finally mentioning about master itself acting as client in case of openshift on rhel [Divya] I have updated the documentation based on our discussion. comment # 2 architecture diagram doesn't really represent RHGS converged with openshift. A node should have gluster and applicattion pods residing together and multiple such nodes should be shown to form a cluster. current diagram doesn't depict the converged solution correctly. [Divya] It was decided that we retain the architecture diagram as is for now. comment # 3: section 3.2.1.2 --> Information about 'subscribe to openshift repo' is provided when a client is configured separately, same will have to be done for master as well. the way the steps unfold is misleading. [Divya] Further to our discussion, no changes are required for this comment. comment # 4: Either a note or link to the setup of openshift cluster has to be provided before section 3.2.3 [Divya] I have added a link to the openshift documentation. comment # 5: After, Execute the following command to reload the iptables: # systemctl reload iptables A validation step can be added, iptables -L on all nodes to validate if actually iptables have been updated. [Divya] I have added the iptalbes verification command. comment # 6: section 3.2.3, last point, Ensure the nodes have the valid ports opened for communicating with Red Hat Gluster Storage. See the iptables configuration task in Step 1 of Section 4.1, “Preparing the Red Hat OpenShift Container Platform Cluster” for more information. The link doesn't provide any information on ports that are opened. [Divya] Further to our discussion, I have deleted this bullet point as it is no longer applicable. comment # 7: section 4.1, note under step 3 says, If a router already exists, proceed to Step 5. It would be good if appropriate command is provided to see if router is already configured. [Divya] Bhavana has updated this information as part of comment # 11. comment # 8: section 4.1, under step 3 --> while step 1&2 has to be run on client, step 3&4 has to be run on master node. [Divya] Updated the steps in this section accordingly. comment # 9: section 4.1, under step 3 --> step 3 & 4 -> what do we do if we have more than 1 master node? Information on whether to edit on all master nodes need to be updated [Divya] Updated the steps in this section accordingly. comment # 10: section 4.1, step 3 has further 4 sub-steps. they are also numbered 1,2,3,4. Instead (i), (ii), (iii) and (iv) can be used. This change has to be done all over the document. [Divya] It is formatted correctly as orderedlist in the xml file but the numbering style is rendered as per Red Hat CCS style sheet. I will bring this up with the CCS team and get this fixed. Please keep in mind that this build contains only comments 1 to 10 updates and Bhavana's updates are not yet merged to the master.
Thanks Ashiq and Karthik. The following link has all the changes: http://ccs-jenkins.gsslab.brq.redhat.com:8080/job/doc-Red_Hat_Gluster_Storage-3.4-Container_Native_Storage_with_OpenShift_Platform-branch-master/lastSuccessfulBuild/artifact/tmp/en-US/html-single/index.html
comments: 1) please revisit comment # 3. 'Subscribe to the OpenShift Container Platform 3.4 repository' still seems to be under section 3.2.1.2. This section is common for 3.2.1.1 and 3.2.1.2 2) comment # 8: I see that you've made the changes for 1 & 2 too. Please note that 1&2 has to be run on client, while 3&4 has to be run on master we can talk about it if you need any clarification. 3) comment # 9, can you point out what changes has been made? I don't see the change. other changes looks fine.
(In reply to krishnaram Karthick from comment #14) > comments: > > 1) please revisit comment # 3. > > 'Subscribe to the OpenShift Container Platform 3.4 repository' still seems > to be under section 3.2.1.2. This section is common for 3.2.1.1 and 3.2.1.2 > According to our previous discussion, no changes are required for this comment. Hence, no changes made. > 2) comment # 8: > > I see that you've made the changes for 1 & 2 too. Please note that 1&2 has > to be run on client, while 3&4 has to be run on master > Updated the doc as per our discussion: http://ccs-jenkins.gsslab.brq.redhat.com:8080/job/doc-Red_Hat_Gluster_Storage-3.4-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-Setting_the_environment-Preparing_RHOE > we can talk about it if you need any clarification. > > 3) comment # 9, can you point out what changes has been made? I don't see > the change. Your comments 8 and 9 are for the same section.
changes made looks good to me, moving the bug to verified.