Description of problem: ======================= In the WebAdmin document, cluster expansion is not documented. We should have a new chapter describing "Cluster Expansion" something similar to the node expansion we have for "RHGS Admin Guide". It should cover below: 1. Unmanage your existing cluster 2. Expand your storage nodes 3. Re-import your cluster
Added new chapter in the Monitoring Guide based on the information provided in Comment 8. Chapter 3. Cluster Expansion https://access.qa.redhat.com/documentation/en-us/red_hat_gluster_storage/3.3/html-single/monitoring_guide/#cluster_expansion Keeping this in MODIFIED state till there is a response to Comment 9.
There's stopping of two services for Gluster server collectd and tendrl-node-agent in chapter '3.3.1. Commands for Stopping Tendrl Services'. But there's another service running on the Gluster server machine - tendrl-monitoring-integration, it should be stopped too.
Sorry the service should be tendrl-gluster-integration not tendrl-monitoring-integration in Comment 24. Something is missing in section '3.3. Reimporting Cluster'. There's sentence See and nothing more. There should be link used for 'See Chapter 2. Import Cluster of this Guide.' in chapter '3.3. Reimporting Cluster'.
(In reply to Lubos Trilety from comment #24) > There's stopping of two services for Gluster server collectd and > tendrl-node-agent in chapter '3.3.1. Commands for Stopping Tendrl Services'. > But there's another service running on the Gluster server machine - > tendrl-monitoring-integration, it should be stopped too. As discussed with Ankush, when you stop the node-agent (onstorage nodes, gluster-integration is stopped as well so it's alright if this is not in the doc)
Verified with latest document. Section 3.3.3 has been removed from the doc. The ordering in chapter 3 has been addressed. Hence marking it as verified
Expanding cluster is by 4 steps via the guide 1. Unmanage the existing cluster from Web Administration. 2. Expand the Gluster storage nodes. 3. Install the Web Administration components via tendrl-ansible. 4. Reimport the cluster in the Web Administration environment. The first is pretty covered in 3.1 sections. - The second is named correctly but content is bad. Because it's about install and expand together. - Moreover the order is wrong, first expand has to be made then install, not vice versa. - And I miss the link to install section in Quick Start Guide for the install part. - - Anyway the install part should be in another unique section e.g. '3.3 Installing Web Administration'
(In reply to Lubos Trilety from comment #34) > Expanding cluster is by 4 steps via the guide > 1. Unmanage the existing cluster from Web Administration. > 2. Expand the Gluster storage nodes. > 3. Install the Web Administration components via tendrl-ansible. > 4. Reimport the cluster in the Web Administration environment. > > The first is pretty covered in 3.1 sections. > > - The second is named correctly but content is bad. Because it's about > install and expand together. > - Moreover the order is wrong, first expand has to be made then install, not > vice versa. > - And I miss the link to install section in Quick Start Guide for the > install part. - - Anyway the install part should be in another unique > section e.g. '3.3 Installing Web Administration' Lubos, I've modified the section. Please take a look. Do we need the start the services on the storage nodes or tendrl-ansible will do that? https://doc-stage.usersys.redhat.com/documentation/en-us/red_hat_gluster_storage/3.3/html-single/monitoring_guide/#expanding_storage_nodes
> Lubos, > > I've modified the section. Please take a look. Do we need the start the > services on the storage nodes or tendrl-ansible will do that? > > https://doc-stage.usersys.redhat.com/documentation/en-us/ > red_hat_gluster_storage/3.3/html-single/monitoring_guide/ > #expanding_storage_nodes tendrl-ansible start services so no need to start them manually. On the other hand nothing serious if it stays in the guide. It doesn't break anything. BTW there's typo: After the cluster in expanded --> After the cluster is expanded One other note, it's not big deal but I really like to see install as another chapter to properly correspond expansion steps. All other observations are fixed.
Hi Team, I've made the required change. Following is the link to the updated content: https://doc-stage.usersys.redhat.com/documentation/en-us/red_hat_gluster_storage/3.3/html-single/monitoring_guide/#expanding_storage_nodes Let me know in case of any concerns.
There are several issues which has to be addressed for the expanding cluster procedure Bug 1526343 and Bug 1526375. The first one could be easily solved by stopping httpd service on server together with all those web administrator services in 3.1.1 Commands for Stopping Tendrl Services. The second could be solved by not uninstalling all tendrl-* and related services on Tendrl server. And if we don't uninstall any package on server that will solve both issues together.
(In reply to Lubos Trilety from comment #38) > There are several issues which has to be addressed for the expanding cluster > procedure > Bug 1526343 and Bug 1526375. > The first one could be easily solved by stopping httpd service on server > together with all those web administrator services in 3.1.1 Commands for > Stopping Tendrl Services. > The second could be solved by not uninstalling all tendrl-* and related > services on Tendrl server. > And if we don't uninstall any package on server that will solve both issues > together. The last sentence is not correct, httpd service has to be reloaded after all. Or, as I suggested, stopped before tendrl-ansible re-install.
For BZ: 1526343 =============== httpd issue mentioned in bug 1526343 is not happening if the packages are not uninstalled at tendrl-server. This change (removal of uninstalling packages at tendrl server) is already addressed in the documentation mentioned in comment 37. For BZ: 1526375 =============== This is an engineering BZ and to be addressed in subsequent releases. This issue will also not occur if we do not uninstall the packages which is already corrected in document. I do not think these to be addrresed via document any further as "uninstalling packages at tendrl server is removed"
All demands were fixed. Uninstall of packages on server is not part of the un-manage cluster procedure anymore. Via Comment 40 httpd doesn't have to be reloaded after expand so no need to stop httpd service. However if it happens reload of httpd is enough to fix any issue. Tendrl services doesn't have to be started manually but it doesn't broke anything, so it doesn't matter. The install doesn't have special chapter but the order is correct. That said the guide looks fine.