[7609] in Release_7.7_team
RE: install failure
daemon@ATHENA.MIT.EDU (Pam Nicholas)
Wed Aug 17 09:24:45 2011
From: "Pam Nicholas" <pmn@MIT.EDU>
To: Jonathan D Reed <jdreed@mit.edu>
CC: "release-team@mit.edu" <release-team@mit.edu>
Date: Wed, 17 Aug 2011 09:24:35 -0400
Message-ID: <7813A033ABE64A41BE07B89F0EDFED3423D585DA8A@EXPO22.exchange.mit.edu>
In-Reply-To: <FF2FD9AC-57D2-4B52-9D95-EE7A514ABDA2@mit.edu>
Content-Language: en-US
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Hi jonathan,
I found your email hiding in the Junk folder for some reason.
This problem is strange because I succesfully imaged a small ff 740 last week. And some of the full size 740s are also giving problems. We'll try your suggestions in order to get up and running quickly.
Thanks
Pam
-----Original Message-----
From: Jonathan Reed [mailto:jdreed@MIT.EDU]
Sent: Wednesday, August 17, 2011 9:09 AM
To: Pam Nicholas
Cc: release-team@mit.edu
Subject: Re: install failure
Hi Pam,
We're a bit baffled, since this would imply some fundamental hardware incompatibility between the network-based installer and the hardware. Certainly it works on regular 745s, and without the SFF hardware to test on, it may be difficult to debug this. We have definitely seen issues with other Dell Small Form Factor Optiplexes in the past.
If you're pressed for time, there are a couple of options:
- Install Ubuntu from CD and install the "workstation" metapackage (using the install script from http://debathena.mit.edu/install) instead and set a strong root password. Users won't be able to install any software themselves or use the "kiosk" web browser, but everthing else should work fine.
- Install Ubuntu from CD and install the "cluster" metapackage (again, using the install script. The installer will accept this as an option, even though it's not listed). Most everything should work, though we haven't yet widely tested this option.
-Jon
On Aug 17, 2011, at 8:24 AM, Pam Nicholas wrote:
> Hi
>
> I want to move several Optiplex 740s in the Libraries to DebAthena this week, so can someone please get back to me on this?
>
> Thanks.
> Pam
>
> -----Original Message-----
> From: Pam Nicholas
> Sent: Monday, August 15, 2011 4:20 PM
> To: Jonathan D Reed
> Cc: Geoffrey G Thomas; release-team@mit.edu; Noel Atkins
> Subject: RE: install failure
>
> I was able to successfully image both machines. However, we've run into problems with other small form factor Optiplex 745s.
>
> Attached is a screen shot of where the process hangs. Any ideas? One of the hostnames is lib-lhum-04.
>
> Some additional info:
> BIOS version does not matter
> Resetting BIOS does not fix it
> Ubuntu 11.04 x86_64 can be installed without issue from a CD RHEL can be installed from MIT PXE
>
>
> -Pam
>
> -----Original Message-----
> From: Jonathan Reed [mailto:jdreed@MIT.EDU]
> Sent: Wednesday, July 27, 2011 4:29 PM
> To: Pam Nicholas
> Cc: Geoffrey G Thomas; release-team@mit.edu
> Subject: Re: install failure
>
> That is correct, DNS updates generally take at least 24 hours. However, the installer probably should have warned you about that earlier in the installation. We will look into fixing that.
>
> -Jon
>
> On Jul 27, 2011, at 4:25 PM, Pam Nicholas wrote:
>
>> Ctrl-Alt-F5 told me that "The IP address you selected does not have an associated hostname."
>>
>> I had the hosts/IPs reassigned to this subnet and got an email that the work was done, but apparently the DNS records weren't yet updated when I ran the installer.
>>
>> Running it again now.
>>
>> Pam
>> -----Original Message-----
>> From: Jonathan Reed [mailto:jdreed@MIT.EDU]
>> Sent: Wednesday, July 27, 2011 3:01 PM
>> To: Pam Nicholas
>> Cc: Geoffrey G Thomas; release-team@mit.edu
>> Subject: Re: install failure
>>
>> Press Ctrl-Alt-F5 to switch to the installer terminal and let us know if you see any error messages or anything unusual on the screen?
>>
>> -Jon
>>
>> On Jul 27, 2011, at 11:59 AM, Pam Nicholas wrote:
>>
>>> I re-ran the installer on these two machines using 18.42.4.98 and
>>> 18.42.4.99 I kicked off the install at 5:27, and when I came in this morning, they were both hung at the screen that says "this installation can take 3 hours etc."
>>>
>>> Any ideas?
>>>
>>> -----Original Message-----
>>> From: Jonathan Reed [mailto:jdreed@MIT.EDU]
>>> Sent: Tuesday, July 26, 2011 11:04 AM
>>> To: Pam Nicholas
>>> Cc: Geoffrey G Thomas; release-team@mit.edu
>>> Subject: Re: install failure
>>>
>>> Ah, that would explain the problem. The machine needs to be installed on the network where it will be used. Additionally, the cluster environment does not support DHCP, only static IP addresses.
>>>
>>> If you wish to use the debathena-workstation metapackage, you can install stock Ubuntu linux (from an Ubuntu CD), and then follow the instructions at http://debathena.mit.edu/install to install Debathena. That should work for a machine which has been registered for DHCP. However, note that debathena-workstation machines will require manual intervention every August to upgrade to the latest release. They also have a slightly different configuration from workstations in the Athena clusters.
>>>
>>> -Jon
>>>
>>> On Jul 26, 2011, at 10:58 AM, Pam Nicholas wrote:
>>>
>>>> Hi Jon,
>>>>
>>>> I was actually running the installer from my office in E25, I assumed since the hostnames/mac addresses were set up for dhcp that should work. Is that not the case?
>>>>
>>>>
>>>> Pam
>>>> -----Original Message-----
>>>> From: Jonathan Reed [mailto:jdreed@MIT.EDU]
>>>> Sent: Tuesday, July 26, 2011 10:57 AM
>>>> To: Pam Nicholas
>>>> Cc: Geoffrey G Thomas; release-team@mit.edu
>>>> Subject: Re: install failure
>>>>
>>>> Hi Pam,
>>>>
>>>> I just tested the installer and it was able to successfully configure the correct IP addresses for those hostnames. Obviously, I can't test any further than that since I'm not in E53 or building 14, respectively. I would normally suspect some local network problem, but obviously if you were able to boot over the network, that's not the case. I can meet you in Dewey Library at some point this week and take a look in person, if you send me some times when you're available. Something strange may be going on on the E53 network, as the cluster services team also had a problem installing another machine in Dewey (dew-ref-1), with a similar failure.
>>>>
>>>> -Jon
>>>>
>>>> On Jul 26, 2011, at 10:37 AM, Pam Nicholas wrote:
>>>>
>>>>> I entered hostnames -- lib-adew-01 and lib-pre-01
>>>>>
>>>>> -----Original Message-----
>>>>> From: Jonathan Reed [mailto:jdreed@MIT.EDU]
>>>>> Sent: Tuesday, July 26, 2011 5:22 AM
>>>>> To: Pam Nicholas
>>>>> Cc: Geoffrey G Thomas; release-team@mit.edu
>>>>> Subject: Re: install failure
>>>>>
>>>>> Hi Pam,
>>>>>
>>>>> This indicates some error in the networking configuration at installation time. When the installer prompted for an IP address after you selected a cluster installation, can you let us know what value you entered?
>>>>>
>>>>> -Jon
>>>>>
>>>>> On Jul 25, 2011, at 4:03 PM, Pam Nicholas wrote:
>>>>>
>>>>>> Hi,
>>>>>>
>>>>>> Any ideas on this?
>>>>>>
>>>>>> Thanks.
>>>>>> Pam
>>>>>>
>>>>>> -----Original Message-----
>>>>>> From: Pam Nicholas
>>>>>> Sent: Friday, July 22, 2011 12:02 PM
>>>>>> To: Geoffrey Thomas
>>>>>> Subject: RE: install failure
>>>>>>
>>>>>>
>>>>>>
>>>>>> chmod: kexec: No such file or directory
>>>>>>
>>>>>> and then later
>>>>>>
>>>>>> debathena/installer.sh: line 312: ./kexec: not found
>>>>>>
>>>>>> Nothing displays when I type ls -l /h
>>>>>> ________________________________________
>>>>>> From: Geoffrey Thomas [geofft@MIT.EDU]
>>>>>> Sent: Thursday, July 21, 2011 12:24 PM
>>>>>> To: Pam Nicholas
>>>>>> Cc: release-team@mit.edu
>>>>>> Subject: Re: install failure
>>>>>>
>>>>>> Hi Pam,
>>>>>>
>>>>>> Were there errors earlier on the screen about downloading the kexec file?
>>>>>> (You can use Shift-PgUp and Shift-PgDn to scroll back.)
>>>>>>
>>>>>> Can you run "ls -l /h" at the debugging shell and let us know what
>>>>>> the output is?
>>>>>>
>>>>>> Thanks,
>>>>>> --
>>>>>> Geoffrey Thomas
>>>>>> IS&T Athena Release Team
>>>>>> release-team@mit.edu
>>>>>>
>>>>>> On Thu, 21 Jul 2011, Pam Nicholas wrote:
>>>>>>
>>>>>>>
>>>>>>> I am trying to install debathena, the cluster version, on
>>>>>>> Optiplex 740 and 745 computers. Both failed with the message
>>>>>>>
>>>>>>> Line312 ./kexec: not found
>>>>>>>
>>>>>>> Secondary installed failed
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> Any advice?
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> Thanks
>>>>>>>
>>>>>>> Pam
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> ---
>>>>>>>
>>>>>>> Pam Nicholas
>>>>>>> Computer Support Manager
>>>>>>> MIT Libraries, Rm E25-131
>>>>>>> Cambridge, MA 02139
>>>>>>> 617-253-1612
>>>>>>> pmn@mit.edu
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>
>>>>
>>>
>>
>
> <debathena fail.jpg>