Bug 1770030
Summary: | [4.4.0-5] after deploy of HE the defined fqdn on the host changed to localhost.localdomain | ||||||
---|---|---|---|---|---|---|---|
Product: | [oVirt] ovirt-ansible-collection | Reporter: | Kobi Hakimi <khakimi> | ||||
Component: | hosted-engine-setup | Assignee: | Dominik Holler <dholler> | ||||
Status: | CLOSED NOTABUG | QA Contact: | Nikolai Sednev <nsednev> | ||||
Severity: | high | Docs Contact: | |||||
Priority: | high | ||||||
Version: | unspecified | CC: | bugs, dholler, dron, jmacku, lsvaty, mavital, mburman, mperina, stirabos | ||||
Target Milestone: | ovirt-4.4.0 | Keywords: | AutomationBlocker | ||||
Target Release: | 1.0.34 | Flags: | sbonazzo:
ovirt-4.4?
|
||||
Hardware: | Unspecified | ||||||
OS: | Unspecified | ||||||
Whiteboard: | |||||||
Fixed In Version: | ovirt-ansible-hosted-engine-setup-1.0.34 | Doc Type: | If docs needed, set a value | ||||
Doc Text: | Story Points: | --- | |||||
Clone Of: | Environment: | ||||||
Last Closed: | 2019-12-23 08:05:10 UTC | Type: | Bug | ||||
Regression: | --- | Mount Type: | --- | ||||
Documentation: | --- | CRM: | |||||
Verified Versions: | Category: | --- | |||||
oVirt Team: | Network | RHEL 7.3 requirements from Atomic Host: | |||||
Cloudforms Team: | --- | Target Upstream Version: | |||||
Embargoed: | |||||||
Bug Depends On: | 1779182 | ||||||
Bug Blocks: | 1701490 | ||||||
Attachments: |
|
host deploy flow doesn't alter the hostname in any way How did you set host name? just plain hostname command or hostnamectl? we reprovision the host with foreman and when I run hostnamectl before the deploy I got: [root@caracal01 ~]# hostnamectl Static hostname: localhost.localdomain Transient hostname: caracal01.lab.eng.tlv2.redhat.com Icon name: computer-server Chassis: server Machine ID: f7cff39c87324de9a5e42ceb60139e7f Boot ID: c16d8f55216f4a38ab39779c052be6cb Operating System: Red Hat Enterprise Linux 8.1 (Ootpa) CPE OS Name: cpe:/o:redhat:enterprise_linux:8.1:GA Kernel: Linux 4.18.0-147.el8.x86_64 Architecture: x86-64 [root@caracal01 ~]# after hosted engine deploy: [root@caracal01 ~]# hostnamectl Static hostname: localhost.localdomain Icon name: computer-server Chassis: server Machine ID: f7cff39c87324de9a5e42ceb60139e7f Boot ID: c16d8f55216f4a38ab39779c052be6cb Operating System: Red Hat Enterprise Linux 8.1 (Ootpa) CPE OS Name: cpe:/o:redhat:enterprise_linux:8.1:GA Kernel: Linux 4.18.0-147.el8.x86_64 Architecture: x86-64 [root@caracal01 ~]# So I think the transient hostnames set/updated by DHCP in the beginning according to @simone maybe as a side effect of this issue the hosted engine deploy failed with the following error: 16:42:38 TASK [ovirt.hosted_engine_setup : Shutdown local VM] *************************** 16:42:39 An exception occurred during task execution. To see the full traceback, use -vvv. The error was: libvirt.libvirtError: unable to connect to server at 'lynx22.lab.eng.tlv2.redhat.com:16514': Connection refused 16:42:39 fatal: [lynx22.lab.eng.tlv2.redhat.com]: FAILED! => {"changed": false, "msg": "unable to connect to server at 'lynx22.lab.eng.tlv2.redhat.com:16514': Connection refused"} in code it failed in: https://github.com/oVirt/ovirt-ansible-hosted-engine-setup/blob/master/tasks/create_target_vm/03_hosted_engine_final_tasks.yml#L90 as you can see in: https://rhv-jenkins.rhev-ci-vms.eng.rdu2.redhat.com/job/rhv-ge-deploy-4.4/123/console it seems the hosted engine deployment failed early, libvirt tls socket activation failed on port 16514, looks similar to https://bugzilla.redhat.com/show_bug.cgi?id=1752837 we need to to work with libvirt > 5.6.0-6 after upgrading to package from https://brewweb.engineering.redhat.com/brew/buildinfo?buildID=1013923 libvirt-5.6.0-6.module+el8.1.0+4244+9aa4e6bb.x86_64 tls socket activation is working: virsh -c qemu+tls://lynx01.lab.eng.tlv2.redhat.com:16514/system list Id Name State ----------------------------------- 1 HostedEngineLocal running Seems fixed by https://access.redhat.com/errata/RHBA-2019:3723. Moving to QA to verify it. Failed in libvirt-5.6.0-6.1.module+el8.1.0+4754+8d38b36b.x86_64 12:41:03 < tiraboschi> dfodor: sbonazzo: the same even with libvirt-5.6.0-6.1.module+el8.1.0+4754+8d38b36b.x86_64 The following combination libvirt 5.6.0-6.1 and ovirt-ansible-hosted-engine-setup-1.0.34-1.el8ev.noarch.rpm fix this issue. The issue seems to reproduce without HE at all. Add host to rhv 4.4.0-6 over hostname, the hostname is overridden and it's now localhost.localdomain Should i open a separate bug? (In reply to Michael Burman from comment #10) > The issue seems to reproduce without HE at all. > Add host to rhv 4.4.0-6 over hostname, the hostname is overridden and it's > now localhost.localdomain > Should i open a separate bug? Yes please, cam you also run hostnamectl, before and after the host deployment. (In reply to Evgeny Slutsky from comment #11) > (In reply to Michael Burman from comment #10) > > The issue seems to reproduce without HE at all. > > Add host to rhv 4.4.0-6 over hostname, the hostname is overridden and it's > > now localhost.localdomain > > Should i open a separate bug? > > Yes please, > cam you also run hostnamectl, before and after the host deployment. New bug for the non HE scenario. All info is there. tnx (In reply to Michael Burman from comment #12) > (In reply to Evgeny Slutsky from comment #11) > > (In reply to Michael Burman from comment #10) > > > The issue seems to reproduce without HE at all. > > > Add host to rhv 4.4.0-6 over hostname, the hostname is overridden and it's > > > now localhost.localdomain > > > Should i open a separate bug? > > > > Yes please, > > cam you also run hostnamectl, before and after the host deployment. > New bug for the non HE scenario. All info is there. tnx BZ 1779182 Would be setting the static host name to the transient host name a workaround for you, until nmstate/NetworkManager is enabled by default? (In reply to Dominik Holler from comment #14) > Would be setting the static host name to the transient host name a > workaround for you, until nmstate/NetworkManager is enabled by default? after https://github.com/oVirt/ovirt-ansible-hosted-engine-setup/pull/275 merged and we added the following task before deploy: shell: hostnamectl set-hostname {{ inventory_hostname }} I don't have this issue anymore. Kobi, can you please check if the static hostname is set in the kickstart file used to deploy the host? The kickstart file may contain a line like network --bootproto dhcp --hostname xxx.xx.xxx.tlv.redhat.com --device=xxx:xx:xx:xx:xx:xx If such or a similar line is not in the kickstart file (or this line is commented), how is the static hostname set? I checked all our kickstart templates used to provision host based on RHEL 8.1 (including rhvh-4.4). We do have this line in all our kickstart files and this line is not commented out. (In reply to Jiri Macku from comment #17) > I checked all our kickstart templates used to provision host based on RHEL > 8.1 (including rhvh-4.4). > > We do have this line in all our kickstart files and this line is not > commented out. I understood that this answer is a generic one, not about the affected host. Can you please check if the issue reproduces if the /etc/hosts file contains the expected hostname before starting the RHV deployment? (In reply to Dominik Holler from comment #18) > (In reply to Jiri Macku from comment #17) > > I checked all our kickstart templates used to provision host based on RHEL > > 8.1 (including rhvh-4.4). > > > > We do have this line in all our kickstart files and this line is not > > commented out. > > I understood that this answer is a generic one, not about the affected host. > Can you please check if the issue reproduces if the /etc/hosts file contains should be /etc/hostname > the expected hostname before starting the RHV deployment? after mburman and I checked it, we found out that indeed the following line: network --bootproto dhcp --hostname xxx.xx.xxx.tlv.redhat.com --device=xxx:xx:xx:xx:xx:xx was comment out in rhel-8.1 kickstart file. the kickstart fixed and according to mburman we don't see this issue anymore(regular env) Kobi, can this bug be closed now? (In reply to Dominik Holler from comment #21) > Kobi, can this bug be closed now? Yep, from our side, you can close it. Thanks!! Kobi, thanks for checking. |
Created attachment 1633813 [details] host logs Description of problem: [4.4.0-4] after deploy of HE the defined fqdn on the host changed to localhost.localdomain Version-Release number of selected component (if applicable): ovirt-ansible-engine-setup-1.1.9-1.el8ev.noarch python3-ovirt-setup-lib-1.3.0-0.0.master.20190419120545.gitfbe1cbd.el8ev.noarch ovirt-hosted-engine-setup-2.4.0-0.1.master.20191104160243.git0c51343.el8ev.noarch How reproducible: 100% Steps to Reproduce: 1. run the command: "hostname" to make sure your host fqdn defined as expected. 2. deploy the HE over ISCSI Actual results: the defined fqdn on the host changed to localhost.localdomain Expected results: leave the fqdn as it was Additional info: also happened on deploy of HE over NFS