VMware Cloud Community
GangFeng
Contributor
Contributor

Why can not install STAF on ESXi5 by using workbench2.0 installation helper?

Hi, experts

When I'm trying run a certification for ESXi5, I can not install STAF on ESXi5.0.

Acturally, the installation process seems completed, but STAF ping failed, so installation helper report errors which says STAF installation failed.

I'm pretty sure I turned firewall off on ESXi5.0, I can ssh to this ESX5.0 from the CentOS ( workbench2.0 ).

I attached the screenshot of the fail info.

Wish you experts can help me out. Thanks in advance.

Best regards

--Feng

Tags (3)
Reply
0 Kudos
16 Replies
GangFeng
Contributor
Contributor

Is there any one already begin to use wb2.0 to do ESXi5 certification? Any one meet the same issue?

Reply
0 Kudos
GangFeng
Contributor
Contributor

It turns out that STAFProc failed to run on workbench because of the dns resolving issue.

By adding ip address and hostname into /etc/hosts fix this problem.

Reply
0 Kudos
orevueltas
Contributor
Contributor

Hey GangFeng,

I'm too running the vSphere 5 Cert and I'm running into the same issue.

Which host/ip did you added to the /etc/hosts file?

mine looks

127.0.0.1     localhost     workbench
172.32.16.54     workbench.solutions.drobo.com     workbench

Did you added to the workbenck vm or the ESXi host?

Regards

Omar R

Reply
0 Kudos
VladL
Enthusiast
Enthusiast

Hi all,

most STAF deployment issues from workbench to ESXi caused by :

1. WB and ESXi MUST be reachable on same network, before installing STAF open shell on WB and ping ESXi on trusted network interface.

2. [ESXi] /etc/resolv.conf

     -     ensure you have ONE entry for "nameserver" and "search" - multiple serch domains cause trouble for STAF framework.

3. [ ESXi & WB ] /etc/hosts

     -     ensure both, WB and target ESXi, appear in this file on the two machines, example :

10.11.12.13   some-host-name.some-domain-name.com
10.11.14.15   workbench
Hope this helps - good Luck,
Vlad.
Reply
0 Kudos
orevueltas
Contributor
Contributor

In the WorkBench admin guide it states that you can execute the     application from a remote host using X forwarding using the command

ssh -X -l vmware root@172.32.16.54 vmwb

...if you are executing this from a Linux or Mac host, we see the     Workbech application on a X11 window. However, when trying to     install STAF on a ESX/i host you then get the error that it failed     to ping and all the java output

STAF_error.jpg

What worked for me to get around this issue is to:

ssh -X vmware@172.32.16.54

...at the command prompt

[root@workbench ~]# gnome-session

...this will bring the entire gnome GUI of the workbench(CentOS)     onto your host.

Now open the Workbench application from the Desktop, and try to     install STAF on the ESX/i host and you shall not see this error     anymore.

Also, if you are using a Mac, you can enable XDMCP within CentOS/Workbench. All you need to do is modify a file

[root@workbench /]# vi /etc/gdm/custom.conf

...find

[xdmcp]

...and add

Enable=true

...beneath it so it would look like this

[xdmcp]
Enable=true

...save the file and either reboot the Workbench VM or enter into runlevel 5 (init 5).

Once this is complete, you can use OS X Xquartz command

SE-Admins-Mac-Pro:~ orevueltas$ Xquartz -query 172.32.16.54

Reply
0 Kudos
davidturchak
Contributor
Contributor

Vlad, I have the same issue with STAF installation. Both hosts added to hosts file and successfully ping each other .... this is not a connectivity issue for sure.

I'm using 3.0.1 WB and ESXi 5.5. (firewall disabled and unloaded ) Any way to debug this behavior?

-David

Reply
0 Kudos
VladL
Enthusiast
Enthusiast

Hi David,

does your ESXi host on which you attempt to install STAF is currently claimed by some vCenter ?

If it does - remove it from vCenter inventory, run #services.sh restart from ESXi host shell (verify successful start/stop cycle),  and attempt to install again.

Reply
0 Kudos
davidturchak
Contributor
Contributor

Hi Vlad,

Thanks for your fast response. My ESXi is not claimed in vCenter, it's a fresh installed host .

I see that staf 3.4.1 is installed in /opt/staf-3.4.1/ , but it's unable to start :

~ # /etc/init.d/staf start

~ # /etc/init.d/staf status

STAF is not running

Where can I find staf log file?

-David

Reply
0 Kudos
VladL
Enthusiast
Enthusiast

Do you have /opt/staf-<ver_num>/staf.out file ?

In general, I would try to enable debug trace in "/opt/staf-3.4.1/bin/STAF.cfg" -> trace enable tracepoints "error deprecated debug"

Reference : http://staf.sourceforge.net/current/STAFUG.htm#HDRTRACEPOINTREF

Reply
0 Kudos
davidturchak
Contributor
Contributor

Looks like some lib is missing :

/opt/staf-3.4.1/bin/STAFProc: error while loading shared libraries: libssl.so.0.9.8: cannot open shared object file: No such file or directory

/opt/staf-3.4.1/bin/STAFProc: error while loading shared libraries: libssl.so.0.9.8: cannot open shared object file: No such file or directory

/opt/staf-3.4.1/bin/STAFProc: error while loading shared libraries: libssl.so.0.9.8: cannot open shared object file: No such file or directory

/opt/staf-3.4.1/bin/STAFProc: error while loading shared libraries: libssl.so.0.9.8: cannot open shared object file: No such file or directory

/opt/staf-3.4.1/staf.out (END)

Who is responsible to provide this file (ESXi installer or installation helper/ staf) ?

Reply
0 Kudos
VladL
Enthusiast
Enthusiast

So basically STAF installation didn't complete from Installation helper via WorkBench ?

Are these prints after adding "debug" trace ?

Is libssl.so.0.9.8 present at /opt/staf-3.4.1/lib/ ?

Reply
0 Kudos
davidturchak
Contributor
Contributor

From progress bar in the bottom of Eclipse it looks like installation is complete 100% , but then failed with staf ping exception.

I didn't add "debug" trace .

There is no /opt/staf-3.4.1/lib/libssl.so.9.8 file present:

/opt/staf-3.4.1/lib # ls -ltr

total 5800

-r-xr-xr-x    1 root     root        175766 Feb 24  2012 libSTAFTCP.so

-r-xr-xr-x    1 root     root        107886 Feb 24  2012 libSTAFReg.so

-r-xr-xr-x    1 root     root        315070 Feb 24  2012 libSTAFPool.so

-r-xr-xr-x    1 root     root        206276 Feb 24  2012 libSTAFMon.so

-r-xr-xr-x    1 root     root        384961 Feb 24  2012 libSTAFLog.so

-r-xr-xr-x    1 root     root        127750 Feb 24  2012 libSTAFLIPC.so

-r-xr-xr-x    1 root     root        121422 Feb 24  2012 libSTAFEXECPROXY.so

-r-xr-xr-x    1 root     root        183599 Feb 24  2012 libSTAFDeviceService.so

-r-xr-xr-x    1 root     root        106205 Feb 24  2012 libSTAFDSLS.so

-r-xr-xr-x    1 root     root       4196152 Feb 24  2012 libSTAF.so

Reply
0 Kudos
VladL
Enthusiast
Enthusiast

In that case, suggest to verify the following firewall rules applied, and persistent after reboot ( try install STAF afterwards again ) :

Default Action: PASS

Enabled: false

Reply
0 Kudos
davidturchak
Contributor
Contributor

:smileyconfused: Same behavior with settings below:

~ # esxcli network firewall get

   Default Action: PASS

   Enabled: false

   Loaded: true

Reply
0 Kudos
davidturchak
Contributor
Contributor

Vlad, I finally solved the issue by manually copy of libssl.so.0.9.8 and libcrypto.so.0.9.8 to my esx. Now staf is running.

Thank you very much for directing me to the right direction Smiley Wink

-David

Reply
0 Kudos
VladL
Enthusiast
Enthusiast

Glad work-around found Smiley Happy

Good luck and best regards.

Reply
0 Kudos