Bug 1643730
Summary: | [upgrade] Post upgrade, new options are not available in virt profile | |||
---|---|---|---|---|
Product: | [Red Hat Storage] Red Hat Gluster Storage | Reporter: | SATHEESARAN <sasundar> | |
Component: | rhhi | Assignee: | Sahina Bose <sabose> | |
Status: | CLOSED CURRENTRELEASE | QA Contact: | SATHEESARAN <sasundar> | |
Severity: | high | Docs Contact: | ||
Priority: | high | |||
Version: | rhhiv-1.5 | CC: | guillaume.pavese, lbailey, rhs-bugs, sabose, sankarshan, sasundar | |
Target Milestone: | --- | Keywords: | Tracking | |
Target Release: | RHHI-V 1.5.z Async | |||
Hardware: | x86_64 | |||
OS: | Linux | |||
Whiteboard: | ||||
Fixed In Version: | imgbased-1.1.2 | Doc Type: | Known Issue | |
Doc Text: |
When a virtualization host is upgraded to Red Hat Hyperconverged infrastructure for Virtualization 1.5, the performance profile for the virt group is not updated with parameters added during development. This means that upgraded deployments are not optimally tuned for most use cases. To work around this issue, add the following options to the /var/lib/glusterd/groups/virt file on each server, or manually set these options on your upgraded volumes.
cluster.choose-local=off
client.event-threads=4
server.event-threads=4
performance.client-io-threads=on
|
Story Points: | --- | |
Clone Of: | ||||
: | 1643733 (view as bug list) | Environment: | ||
Last Closed: | 2019-05-20 04:54:49 UTC | Type: | Bug | |
Regression: | --- | Mount Type: | --- | |
Documentation: | --- | CRM: | ||
Verified Versions: | Category: | --- | ||
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | ||
Cloudforms Team: | --- | Target Upstream Version: | ||
Embargoed: | ||||
Bug Depends On: | 1643733 | |||
Bug Blocks: | 1548985 |
Description
SATHEESARAN
2018-10-28 01:13:29 UTC
[root@ ~]# imgbase w You are on rhvh-4.2.7.3-0.20181026.0+1 [root@ ~]# rpm -qa | grep glusterfs-server glusterfs-server-3.12.2-25.el7rhgs.x86_64 [root@ ~]# rpm -qf /var/lib/glusterd/groups/virt glusterfs-server-3.12.2-25.el7rhgs.x86_64 [root@ ~]# cat /var/lib/glusterd/groups/virt performance.quick-read=off performance.read-ahead=off performance.io-cache=off performance.low-prio-threads=32 network.remote-dio=enable cluster.eager-lock=enable cluster.quorum-type=auto cluster.server-quorum-type=server cluster.data-self-heal-algorithm=full cluster.locking-scheme=granular cluster.shd-max-threads=8 cluster.shd-wait-qlength=10000 features.shard=on user.cifs=off Note that the new options are missing in this virt profile I upgraded the gluster rpms from glusterfs-3.8.4-54.15 to 3.12.2-25 The contents of virt profile were updated. I suspect this to be the problem with RHVH image update. Will just try to understand it with 2 tests Test1 1. Install RHVH 4.1.11 2. Check for contents 3. Do image update from CLI ( # yum update ) 4. Check for contents of virt profile 5. Reboot 6. Check for contents of virt profile Test2 1. Install RHVH 4.1.11 2. Check for contents 3. Remove virt profile 3. Do image update from CLI ( # yum update ) 4. Check for contents of virt profile 5. Reboot 6. Check for contents of virt profile I will update the results for reference This looks like the RHVH related issue. I have tested with single node upgrade as mentioned in comment2 Test1 ----- 1. Created a VM and installed it with RHVH 4.2.6 ( which had glusterfs-3.8.4-54.15.el7rhgs ) 2. Updated to RHVH 4.2.7 image ( redhat-virtuaization-host-image-update ) 3. After update and reboot, I observed the contents of virt profile remained the same. Test2 ------- 1. Created a VM and installed it with RHVH 4.2.6 ( which had glusterfs-3.8.4-54.15.el7rhgs ) 2. Removed the virt profile ( rm -rf /var/lib/glusterd/groups/virt ) 2. Updated to RHVH 4.2.7 image ( redhat-virtuaization-host-image-update ) 3. After updating, I observed that the virt profile is created new and the contents of virt profile had the recently added params. One possible solution to tackle this problem is to call this issue as known_issue. Ask users to enable the options on the volume for performance improvements. The missing perf volume options are: cluster.choose-local=off client.event-threads=4 server.event-threads=4 performance.client-io-threads=on They can achieve it using gdeploy too: [hosts] host1 [volume1] action=set volname=engine key=cluster.choose-local,client.event-threads,server.event-threads,performance.client-io-threads value=off,4,4,on [volume2] action=set volname=vmstore key=cluster.choose-local,client.event-threads,server.event-threads,performance.client-io-threads value=off,4,4,on [volume3] action=set volname=data key=cluster.choose-local,client.event-threads,server.event-threads,performance.client-io-threads value=off,4,4,on Sas, This will need to be documented for existing volumes too - to manually set these options? Laura, do you think this should be added as a note in the guide as well? This issue is fixed with imgbased-1.1.2 as I could see that from the dependent bug, and moving this bug ON_QA. @Sahina, could you provide the devel_ack for this bug ? Verified after updating to RHV 4.2.8 RHVH, and the new virt profile contains all the required volume options |