Re: [ovirt-users] Unexpected protocol mismatch with SPICE HTML5

From: "Michal Skrivanek" <michal.skrivanek(a)redhat.com&gt; To: nicolas(a)devels.es, "Jakub Niedermertl" <jniederm(a)redhat.com&gt; Cc: &quot;users(a)ovirt.org (users(a)ovirt.org)&quot; <users(a)ovirt.org&gt; Sent: Wednesday, June 17, 2015 1:35:04 PM Subject: Re: [ovirt-users] Unexpected protocol mismatch with SPICE HTML5 On Jun 17, 2015, at 09:17 , nicolas(a)devels.es wrote: > Hi Michal, > > El 2015-06-17 07:24, Michal Skrivanek escribió: >> On Jun 17, 2015, at 08:21 , Michal Skrivanek >> <michal.skrivanek(a)redhat.com&gt; wrote: >>> On Jun 16, 2015, at 12:02 , nicolas(a)devels.es wrote: >>>> Hi, >>>> I just made a fresh install of ovirt-engine (v. 3.5.3.1-1) on a CentOS-7 >>>> x86_64 box. My previous installation had issues with the TLS >>>> certificate time of issue, which was successfully fixed in this version >>>> (https://bugzilla.redhat.com/show_bug.cgi?id=1210486). However, when >>>> trying to start the SPICE HTML5 client, all I can see is the grey >>>> square and no activity, and when clicking on "Toggle messages output", >>>> the error message "Unexpected protocol mismatch" is shown. >>>> Things done: >>>> * ovirt-guest-agent installed on the VM >>>> * On client side, updated browser (Firefox) to latest version (38.0) and >>>> /ca.crt has been imported into Firefox certs >>>> Regarding logs on the server side, I can see the following: >>>> ovirt-engine >>>> ============ >>>> 2015-06-16 10:49:29,862 INFO >>>> [org.ovirt.engine.core.bll.SetVmTicketCommand] >>>> (ajp--127.0.0.1-8702-11) [39c79276] Running command: SetVmTicketCommand >>>> internal: false. Entities affected : ID: >>>> c2e32d4f-4c54-4e38-bf66-2627dc6dd4bc Type: VMAction group CONNECT_TO_VM >>>> with role type USER >>>> 2015-06-16 10:49:29,917 INFO >>>> [org.ovirt.engine.core.vdsbroker.vdsbroker.SetVmTicketVDSCommand] >>>> (ajp--127.0.0.1-8702-11) [39c79276] START, >>>> SetVmTicketVDSCommand(HostName = ovirtengine.myorg.es, HostId = >>>> 82fb3c43-3a4b-464e-9077-8e86dea2016b, >>>> vmId=c2e32d4f-4c54-4e38-bf66-2627dc6dd4bc, ticket=TgrRVYoRhC8T, >>>> validTime=120,m userName=user, >>>> userId=5d502bbb-97d6-40b5-abbf-90121cea94d5), log id: 1881c90f >>>> 2015-06-16 10:49:29,937 INFO >>>> [org.ovirt.engine.core.vdsbroker.vdsbroker.SetVmTicketVDSCommand] >>>> (ajp--127.0.0.1-8702-11) [39c79276] FINISH, SetVmTicketVDSCommand, log >>>> id: 1881c90f >>>> 2015-06-16 10:49:29,957 INFO >>>> [org.ovirt.engine.core.dal.dbbroker.auditloghandling.AuditLogDirector] >>>> (ajp--127.0.0.1-8702-11) [39c79276] Correlation ID: 39c79276, Call >>>> Stack: null, Custom Event ID: -1, Message: user user@domain initiated >>>> console session for VM testVM >>>> 2015-06-16 10:49:29,987 WARN >>>> [org.ovirt.engine.core.dal.job.ExecutionMessageDirector] >>>> (ajp--127.0.0.1-8702-11) [7a71a9e4] The message key VmLogon is missing >>>> from bundles/ExecutionMessages >>>> 2015-06-16 10:49:30,064 INFO [org.ovirt.engine.core.bll.VmLogonCommand] >>>> (ajp--127.0.0.1-8702-11) [7a71a9e4] Running command: VmLogonCommand >>>> internal: false. Entities affected : ID: >>>> c2e32d4f-4c54-4e38-bf66-2627dc6dd4bc Type: VMAction group CONNECT_TO_VM >>>> with role type USER >>>> 2015-06-16 10:49:30,114 INFO >>>> [org.ovirt.engine.core.vdsbroker.vdsbroker.VmLogonVDSCommand] >>>> (ajp--127.0.0.1-8702-11) [7a71a9e4] START, VmLogonVDSCommand(HostName = >>>> ovirtengine.myorg.es, HostId = 82fb3c43-3a4b-464e-9077-8e86dea2016b, >>>> vmId=c2e32d4f-4c54-4e38-bf66-2627dc6dd4bc, domain=domain-authz, >>>> password=******, userName=user@domain), log id: 68632b24 >>>> 2015-06-16 10:49:30,120 INFO >>>> [org.ovirt.engine.core.vdsbroker.vdsbroker.VmLogonVDSCommand] >>>> (ajp--127.0.0.1-8702-11) [7a71a9e4] FINISH, VmLogonVDSCommand, log id: >>>> 68632b24 >>>> 2015-06-16 10:49:30,351 INFO >>>> [org.ovirt.engine.core.vdsbroker.gluster.GlusterVolumesListVDSCommand] >>>> (DefaultQuartzScheduler_Worker-2) START, >>>> GlusterVolumesListVDSCommand(HostName = kvmhost.myorg.es, HostId = >>>> 82fb3c43-3a4b-464e-9077-8e86dea2016b), log id: 2a2a439a >>>> 2015-06-16 10:49:30,368 INFO >>>> [org.ovirt.engine.core.vdsbroker.gluster.GlusterVolumesListVDSCommand] >>>> (DefaultQuartzScheduler_Worker-2) FINISH, GlusterVolumesListVDSCommand, >>>> return: {}, log id: 2a2a439a >>>> ovirt-websocket-proxy >>>> ===================== >>>> jun 16 09:56:39 ovirtengine.myorg.es ovirt-websocket-proxy.py[30728]: >>>> 10.X.X.X - - [16/Jun/2015 09:56:39] 10.X.X.X: SSL/TLS (wss://) >>>> WebSocket connection >>>> jun 16 09:56:39 ovirtengine.myorg.es ovirt-websocket-proxy.py[30728]: >>>> 10.X.X.X - - [16/Jun/2015 09:56:39] 10.X.X.X: Version hybi-13, base64: >>>> 'False' >>>> jun 16 09:56:39 ovirtengine.myorg.es ovirt-websocket-proxy.py[30728]: >>>> 10.X.X.X - - [16/Jun/2015 09:56:39] 10.X.X.X: Path: >>>> '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' >>>> Java console >>>> ============ >>>> ">> disconnect" spicehtml5-main.html:86:16 >>>> "<< disconnect" spicehtml5-main.html:90:16 >>>> "ERROR: Error: Unexpected protocol mismatch." >>>> ----------------- >>>> I can't see any other relevant info, so I wonder if something else might >>>> be missing on any of the involved machines. Is there something else to >>>> do? >>> It seems to be an incompatible API change in python-websockify 0.6.0. >>> For now please downgrade to python-websockify-0.5.1-2.el7 >> similar issue with new novnc 0.5.1 has been fixed recently. We're due >> updating both spice-html5 (some small improvements) and adopting >> python-websockify 0.6.0, it will follow in the next couple of >> weeks….if it doesn't make it for 3.6 GA it will be backported > > The downgrade of python-websockify did a substancial progress, since I can > finally see the console :-) However, I have some other issue, as I can see > the console but apparently it won't print anything typed onto the > keyboard. However, the keystrokes indeed seem being sent because if I > close the HTML5 client and reopen it, I see what I typed before I closed > it. The messages output shows something like this: > > Connected to > wss://ovirtengine.myorg.es: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 > > Unimplemented function 108(Inval All Palettes) [ further notices suppressed > ] > > Unimplemented function 102(Display Mark) [ further notices suppressed ] > > FIXME: Unable to interpret bitmap of format: 2 > > 2: Unknown message type 304! > > FIXME: Unable to interpret bitmap of format: 2 > > 2: Unknown message type 304! > > [...the last message repeated x 1000 times...] > > Is this related to the same issue? (tried both on Windows and Linux with > the same result). Hi Nicolás, good…so then it sounds more like a problem of spice-html5…you can try downgrading that too hm..doesn't sound like the web clients are progressing in the right direction :-)…well, that's the reason they are still considered experimental and when possible you should rather use remote-viewer. Jakub was trying that yesterday…did it work for you? Thanks, michal > > Thank you! > > Nicolás > >>> Thansk, >>> michal >>>> Thanks. >>>> Nicolás >>>> _______________________________________________ >>>> Users mailing list >>>> Users(a)ovirt.org >>>> http://lists.ovirt.org/mailman/listinfo/users > _______________________________________________ > Users mailing list > Users(a)ovirt.org > http://lists.ovirt.org/mailman/listinfo/users