Bug 987293
| Summary: | [RHEV+RHS] - On adding RHS server node to Gluster Enabled Cluster on RHEVM 3.2, boot-strapping sets the wrong profile for tuned | |||
|---|---|---|---|---|
| Product: | [oVirt] ovirt-host-deploy | Reporter: | Rejy M Cyriac <rcyriac> | |
| Component: | Plugins.tune | Assignee: | Alon Bar-Lev <alonbl> | |
| Status: | CLOSED CURRENTRELEASE | QA Contact: | SATHEESARAN <sasundar> | |
| Severity: | high | Docs Contact: | ||
| Priority: | unspecified | |||
| Version: | 1.0.0 | CC: | acanan, acathrow, barumuga, bazulay, dougsland, ecohen, grajaiya, iheim, rcyriac, Rhev-m-bugs, sabose, sasundar, srevivo, ssaha, tkatarki, yeylon | |
| Target Milestone: | --- | Flags: | alonbl:
devel_ack+
|
|
| Target Release: | 1.1.0 | |||
| Hardware: | All | |||
| OS: | Linux | |||
| Whiteboard: | gluster | |||
| Fixed In Version: | is10 | Doc Type: | Bug Fix | |
| Doc Text: | Story Points: | --- | ||
| Clone Of: | ||||
| : | 994486 (view as bug list) | Environment: |
virt rhev integration
|
|
| Last Closed: | 2014-01-21 22:17:19 UTC | Type: | Bug | |
| Regression: | --- | Mount Type: | --- | |
| Documentation: | --- | CRM: | ||
| Verified Versions: | Category: | --- | ||
| oVirt Team: | Gluster | RHEL 7.3 requirements from Atomic Host: | ||
| Cloudforms Team: | --- | Target Upstream Version: | ||
| Embargoed: | ||||
| Bug Depends On: | ||||
| Bug Blocks: | 994486 | |||
I am not sure what profile needs to be set in case of gluster only cluster. Options are 1. rhs-high-throughput 2. rhs-virtualization 3. none Note: RHS installation doesn't set any profile by default. Its user choice. is this is a gluster only cluster ? (In reply to Barak from comment #2) > is this is a gluster only cluster ? Yes, the issue was observed on a gluster only cluster. - rejy (rmc) Setting needinfo? as i haven't got a reply for comment #1 Default should be rhs-high-throughput. (In reply to Tushar Katarki from comment #5) > Default should be rhs-high-throughput. Actually now that I re-read the bug, the filer makes it clear in the descriptin that the newly added rhs node should have a tuned profile of rhs-virtualization instead of virtual-host. rhs-virtualization is indeed correct. Please ignore comment #5. Problem here is when adding a node to engine, bootstrap has no clue whether the node is used as vm store or data store. Setting rhs-virtualization will not be correct in case of data store. As tuned profile setting is treated as user's task, I would think that setting it as none (no tuned profile will be set) in bootstrap for rhs is good enough. Please let me know your thoughts before changing the code. After some discussion internally, we are asked to set 'rhs-virtualization' tuned profile (In reply to Bala.FA from comment #8) > After some discussion internally, we are asked to set 'rhs-virtualization' > tuned profile Wasn't that already done[1]? [1] http://gerrit.ovirt.org/#/c/17265/ Yes. comment#8 is a just tracker note. You could move to ON_QA once the build is ready. *** Bug 987077 has been marked as a duplicate of this bug. *** And this and bug#994486 are duplicates. ovirt-host-deploy-1.1.0-0.8.master.el6ev Verified with RHEVM 3.3 IS16 RHS 2.1, RHS 2.0 U5, RHS 2.0 U6 nodes gets "rhs-virtualization" profile when added to gluster-cluster. Closing - RHEV 3.3 Released Closing - RHEV 3.3 Released |
Description of problem: When an RHS server node is added to a Gluster Enabled Cluster in RHEVM 3.2, the boot-strapping process sets the profile 'virtual-host' for tuned. This profile is wrong, as the profile expected to be set for tuned, on an RHS node added to RHEVM, is 'rhs-virtualization' The issue is reproducible irrespective of Gluster Enabled Cluster Compatibility version ------------------------------------------------- 2013-07-23 12:54:11 DEBUG otopi.plugins.ovirt_host_deploy.tune.tuned plugin.executeRaw:347 execute: ('/usr/bin/tuned-adm', 'profile', 'virtual-host'), executable='None', cwd='None', env=None 2013-07-23 12:54:12 DEBUG otopi.plugins.ovirt_host_deploy.tune.tuned plugin.executeRaw:364 execute-result: ('/usr/bin/tuned-adm', 'profile', 'virtual-host'), rc=0 2013-07-23 12:54:12 DEBUG otopi.plugins.ovirt_host_deploy.tune.tuned plugin.execute:412 execute-output: ('/usr/bin/tuned-adm', 'profile', 'virtual-host') stdout: Stopping tuned: [ OK ] Applying ktune sysctl settings: /etc/ktune.d/tunedadm.conf: [ OK ] Calling '/etc/ktune.d/tunedadm.sh start': [ OK ] Applying sysctl settings from /etc/sysctl.conf Applying deadline elevator: sda sdb [ OK ] Starting tuned: [ OK ] 2013-07-23 12:54:12 DEBUG otopi.plugins.ovirt_host_deploy.tune.tuned plugin.execute:417 execute-output: ('/usr/bin/tuned-adm', 'profile', 'virtual-host') stderr: Switching to profile 'virtual-host' 2013-07-23 12:54:12 DEBUG otopi.plugins.otopi.services.rhel rhel.startup:151 set service tuned startup to True ------------------------------------------------- ------------------------------------------------- On RHS 2.0 Update 5: # ls /etc/tune-profiles/ active-profile desktop-powersave functions laptop-battery-powersave rhs-high-throughput server-powersave throughput-performance default enterprise-storage laptop-ac-powersave latency-performance rhs-virtualization spindown-disk # tuned-adm active Current active profile: default Service tuned: enabled, running Service ktune: disabled, stopped On RHS 2.1 # ls /etc/tune-profiles/ active-profile desktop-powersave functions laptop-battery-powersave rhs-high-throughput server-powersave throughput-performance virtual-host default enterprise-storage laptop-ac-powersave latency-performance rhs-virtualization spindown-disk virtual-guest # tuned-adm active Current active profile: virtual-host Service tuned: enabled, running Service ktune: enabled, running ------------------------------------------------- Version-Release number of selected component (if applicable): RHEVM 3.2 - 3.2.1-0.39.el6ev RHS 2.0 Update5 - glusterfs-server-3.3.0.11rhs-1.el6rhs.x86_64 RHS 2.1 - glusterfs-server-3.4.0.12rhs.beta5-2.el6rhs.x86_64 How reproducible: Steps to Reproduce: 1.On RHEVM 3.2, create POSIX compliant FS type Data Center 2.Create Gluster Enabled Cluster 3.Add RHS server node to cluster Actual results: Boot-strapping process sets the wrong profile of 'virtual-host' for tuned, when an RHS server node is added to a Gluster Enabled Cluster on RHEVM 3.2 Expected results: Boot-strapping process should set the expected profile of 'rhs-virtualization' for tuned, when an RHS server node is added to a Gluster Enabled Cluster on RHEVM 3.2 Additional info: When an RHS 2.0 server node is added to Gluster Enabled Cluster of Compatibility version 3.1, the profile setting currently fails, as reported in BZ 987077 . It would succeed if the profile being set is the expected one of 'rhs-virtualization'