From: "Larry Hall (Cygwin)" <reply-to-list-only-lh at cygwin dot com>
To: cygwin at cygwin dot com,
Date: 05/29/2013 03:51 PM
Subject: Re: Postinstall Script Errors With Exit Code 128
Sent by: cygwin-owner at cygwin dot com
On 5/29/2013 3:18 PM, Paul.Nickerson at desknetinc dot com wrote:
So, I think that one of these functions hangs when
cygwin1.dll is present. I'm thinking dump_dodgy_apps.
Possibly. You may have a BLODA problem (
http://cygwin.com/acronyms/#BLODA),
whether known or unknown. If so, getting rid of it will certainly help.
I tried copying cygcheck.exe to other directories and leaving
cygwin1.dll
in bin, and cygcheck does not hang when run from other directories.
Interesting. Maybe there's some Windows permission problems in the
directory you were running from?
I have fixed it. Thank you for your help.
I have a few other AWS EC2 instances. One other was having trouble with
Cygwin install, and none of the rest were. I tried comparing installed
applications, but got nowhere. I also tried installing Windows updates on
the troubled system, and copying the cygwin from a healthy system to the
troubled one. Neither attempt worked. I had looked into folder permissions
before, which all looked fine, and tried installing to another folder and
even another drive, but that hadn't helped.
I then used WinDbg to run C:\cygwin\bin\bash.exe --norc --noprofile
"/etc/postinstall/000-cygwin-post-install.sh", and I think I got this
error: SetContext failed, 0x8007001F. Looking online, it seems that code
is associated with "A device attached to the system is not functioning". I
have already rebooted a bunch of times. I used Belarc Advisor to grab the
virtual hardware version. Belarc Advisor said the System Model / Main
Circuit Board / BIOS on both troubled instances was "Xen HVM domU
3.1.2-92.1.13.el5.", while on the other working instances it was "Xen HVM
domU 3.4.3-2.6.18 09/01/2012".
So, I shut down the instance (I use EBS volumes), created an AMI from it,
created an Instance from the AMI, and started up that instance. This has
the effect of applying a new virtual hardware if available (proven by the
fact that I get a new Instance ID), as well as moving to different
physical hardware. After this, Belarc Advisor reported the good virtual
hardware version. And, I was able to successfully install and use Cygwin
normally. :)
I have attached the output of running that bash.exe command inside WinDbg,
just in case it's useful for anything.
Again, thank you for you help.