Description of problem: At the configure page of "RHEV-M", register rhevh to rhevm1 via password, then register it to another rhevm2 via ip. The rhevm server address is still the previous one although the new one is approved successfully, and in the config file /etc/vdsm-reg/vdsm-reg.cof, it's also the old one. Version-Release number of selected component (if applicable): rhev-hypervisor6-6.4-20121015.1.el6 How reproducible: 100% Steps to Reproduce: 1. Install RHEV-H. 2. Register rhev-h to one rhevm via password. 3. Register rhev-h to another rhevm via ip. Actual results: RHEV-H is still registered in first rhevm. Expected results: RHEV-H register to another rhevm can successful.
Created attachment 633160 [details] vdsm.log
Created attachment 633161 [details] ovirt.log
Created attachment 633162 [details] vdsm-reg.log
bootstrap has been rewritten (rhevm-3.2). vdsm-reg is not used anymore after initial registration, it is disabled. another registration either by TUI or initiated by engine will override all settings.
Created attachment 714423 [details] vdsm.log.rhevh It fails to register to a different rhevm. 1- Assume rhevh is registered already in rhevm-01 2. user ssh admin@rhevh 3. using the admin Text UI the admin changes the management server IP Result: error message appears "Failed downloading certificate" and some trace error on the background see[1]. In addition the management server doesn't change even though the UI shows configuration is successful, and of course we cannot see any hots for approval for rhevm-02. [2] vdsm.log [1] RHEV-M Configuration Failed downloading RHEV-M certificate t:Failed acquiring web certificates Traceback (most recent call last): File "/usr/share/vdsm-reg/deployUtil.py", lineCert vdsm-reg/deployUtil.py", line 1581, in getChainFromSSL File "/usr/lib64/python2.6/site-packages/M2Crypto/SSL/Connection.py", line 185, in connect File "/usr/lib64/python2.6/site-packages/M2Crypto/SSL/Connection.py", line 178, in connect_ssl SSLError: unknown protocol
Created attachment 714445 [details] reg log
Which protocol did you use? http or https? This message is expected if you use http protocol as indeed the certificate cannot be downloaded, but https should be used whenever possible. For the registration, it succeeds: --- MainThread::DEBUG::2013-03-21 21:27:15,563::vdsm-reg-setup::114::root::registerVDS begin. MainThread::DEBUG::2013-03-21 21:27:15,563::vdsm-reg-setup::124::root::registerVDS URI= /RHEVManagerWeb/VdsAutoRegistration.aspx?vds_ip=10.35.102.79&__VIEWSTATE=&vds_unique_id=4C4C4544-0043-4810-8057-C6C04F47354A&vds_name=rose01.qa.lab.tlv.redhat.com&ticket=&port=54321 MainThread::DEBUG::2013-03-21 21:27:15,565::vdsm-reg-setup::142::root::registerVDS failed in HTTPS. Retrying using HTTP. MainThread::DEBUG::2013-03-21 21:27:15,570::vdsm-reg-setup::148::root::registerVDS succeeded using HTTP. MainThread::DEBUG::2013-03-21 21:27:15,570::vdsm-reg-setup::165::root::registerVDS time read: 2013-03-21T21:27:15 MainThread::DEBUG::2013-03-21 21:27:15,570::vdsm-reg-setup::174::root::registerVDS end. --- So engine log is required to see what happened at the other side.
Created attachment 714493 [details] engine.log I used 80 the default port in the text UI is 80 ...
(In reply to comment #14) > Created attachment 714493 [details] > engine.log > > I used 80 > the default port in the text UI is 80 ... Simon, we should change the default port.... right?
Which version of engine do you use? looks like 3.1... while we check 3.2 right? But I see: --- 2013-03-21 23:07:44,661 INFO [org.ovirt.engine.core.register.RegisterServlet] (ajp-/127.0.0.1:8702-12) Succeeded to run RegisterVds. --- However, I also see: --- 2013-03-21 23:11:19,666 ERROR [org.ovirt.engine.core.bll.RegisterVdsQuery] (ajp-/127.0.0.1:8702-7) Query execution failed due to invalid inputs. VDS_STATUS_NOT_VALID_FOR_UPDATE --- Meaning that the node was already added. I need a log for specific reproduce: 1. stop engine. 2. clear logs. 3. start engine. 4. reproduce. 5. stop engine. 6. copy logs.
The default was always the ssl port (8443 in 3.0 and 2.x, 443 in 3.1+) What version of vdsm is in the image being used? Did you register previously using port 80? It remembers the last value used (reads it from vdsm-reg.conf). Does a fresh install show port 80?
(In reply to comment #17) > The default was always the ssl port (8443 in 3.0 and 2.x, 443 in 3.1+) This is what I remember too... and in code it is: --- VDC_HOST_PORT = 443$ --- So I am unsure where this default comes from.
Just tried the rhev-hypervisor-advanced-6.4-20130321.0.iso, I cannot see default port is 80, before and after registration. Please find sequence to reproduce and open a separate bug.
I tested the following sequences each from vanilla iso installation of rhev-hypervisor-advanced-6.4-20130321.0.iso: 1. deploy initiated by engine - PASS 2. registration using HTTPS port 443 - PASS 3. registration using HTTP port 80 - PASS a. I don't see the exception details at the background. b. I do see proper dialog: --- RHEV-M Configuration Failed downloading RHEV-M certificate ---
Created attachment 715444 [details] traceback error in the background attached a screenshot of the error message in the background. When you changed the rhevm (management ip) did you see the node showing in the web admin?
Can you please use the 20130321 version so we can be at sync? Yes, I could see the node. But please sync up and then if it does not work, I will be glad to access your environment.
A bug for the undesired error message at the background was opened: bug#947905. It does not prevent the $subject, nor related.
verified.
3.2 has been released