|
Discussion Forums
|
Thread: Server refused to allocate pty
|
|
|
Replies:
17
-
Pages:
2
[
1
2
| Next
]
-
Last Post:
Nov 6, 2009 2:30 PM
by: agentnate
|
|
|
Posts:
2
Registered:
3/28/08
|
|
|
|
Server refused to allocate pty
Posted:
Mar 28, 2008 7:17 AM PDT
|
|
|
Hi,
I've got an AMI which I've built from the new fedora-8-i386 AMI (ami-f51aff9c). I installed a few bits and pieces on it then did a ec2-bundle-vol on it and uploaded it to s3, registered it, and started the instance and this is all fine (ami-7364811a) and it comes up ok.
I then went in to a running instance of this new AMI and added s3cmd, shut down the daemons I'd added as before, and went through the same process and created another image (priv data changed with snip):
ec2-bundle-vol -c ~/cert.pem -k ~/pk.pem -u snip -p java6-tomcat6-apache2-mysql5-s3cmd
then
ec2-upload-bundle -b snip -m java6-tomcat6-apache2-mysql5-s3cmd.manifest.xml -a snip -s snip
which is ami-f464819d.
After the instance comes up I ssh in and get "Server refused to allocate pty" after the 'Authenticating with' message.
I've compared console outputs on the good one versus the bad and it all looks fine. Searching the web implies fstab being wrong would cause this. Any ideas on how to fix it? A bug in ec2-bundle-vol?
Thanks.
|
|
Posts:
108
Registered:
6/22/06
|
|
|
|
Re: Server refused to allocate pty
Posted:
Mar 28, 2008 7:45 AM PDT
in response to: kieranbaws
|
|
|
Hi Kieran
The most likely cause is probably that the /etc/fstab file in the newly bundle AMI is not correct.
I believe that the bundling tools override the instance /etc/fstab which their own copy which does not work correctly with FC8, to fix this use the '--fstab /etc/fstab'
bundlevol
flag whilst bundling from the original AMI.
The fstab for FC8 should look as follows :
/dev/sda1 / ext3 defaults 1 1
/dev/sda2 /mnt ext3 defaults 0 0
/dev/sda3 swap swap defaults 0 0
none /dev/pts devpts gid=5,mode=620 0 0
none /dev/shm tmpfs defaults 0 0
none /proc proc defaults 0 0
none /sys sysfs defaults 0 0
We'll be sure to fix this in the next release of the bundling tools, please let us know if you have any other issues.
Regards
Marcin
|
|
Posts:
2
Registered:
3/28/08
|
|
|
|
Re: Server refused to allocate pty
Posted:
Mar 28, 2008 8:18 AM PDT
in response to: Marcin@AWS
|
|
|
Great thanks Marcin, that's fixed it
|
|
Posts:
2
Registered:
3/14/06
|
|
|
|
Re: Server refused to allocate pty
Posted:
Mar 28, 2008 1:09 PM PDT
in response to: Marcin@AWS
|
|
|
Hi guys,
Just wanted to say I experienced the same problem bundling that ami, but your fix solved my problems.
Keep up the great work.
Thanks,
Tim
|
|
Posts:
152
Registered:
3/28/08
|
|
|
|
Re: Server refused to allocate pty
Posted:
Mar 30, 2008 7:57 PM PDT
in response to: Marcin@AWS
|
|
|
Thanks so much. Just wanted to add that I encountered this same problem and the solution posted here fixed it.
Can you please reply here to let us know when this is confirmed fixed in the next release of the build tools?
Also I am encountering two related issues with the build process I was hoping someone could chime in on:
1) NOTE: rsync with preservation of extended file attributes failed.
Retrying rsync without attempting to preserve extended file attributes...NOTE: rsync seemed successful but exited with error code 23. This probably means that your version of rsync was built against a kernel with HAVE_LUTIMES defined, although the current kernel was not built with this option enabled. The bundling process will thus ignore the error and continue bundling. If bundling completes successfully, your image should be perfectly usable. We, however, recommend that you install a version of rsync that handles this situation more elegantly.
I saw some other posts about this issue but they are over my head - except the one solution about using rsync from Fedora 4. Anyway is this even necessary? When will this be fixed?
2) Not sure if this is an issue or not but during the bundling process I get this:
Created image.part.33
Generating digests for each part...
Digests generated.
Unable to read instance meta-data for product-codes
Creating bundle manifest...
ec2-bundle-vol complete.
Is the "Unable to read instance meta-data for product-codes" anything to worry about? What causes it and how can I get rid of it?
Thanks!
|
|
Posts:
75
Registered:
1/5/07
|
|
|
|
Re: Server refused to allocate pty
Posted:
Apr 2, 2008 11:12 AM PDT
in response to: rtdev
|
|
|
Both of those notes can be ignored as they are usually non-issues. The tools do a best effort attempt at bundling and print such notes to make you aware of what it tried and how successful it was in its attempt. For instance, while doing rsync, the tools try to preserve security settings on files (something that is usually not a deal-breaker but is needed by SELinux users). If unsuccessful, it retries without attempting to preserve those settings and notifies the user about it.
1. rsync: you can safely ignore this note as long as the bundling tools complete bundling successfully and your instance launches. Eventually, a kernel fix (a xen-capable kernel with full LUTIMES support) and/or rsync fix(rsync which handles link-timestamps gracefully) will become available and this note will disappear.
No need to do anything to continue bundling.
2. Some AMIs (usually, paid AMIs) have functionality that depends on the attached product-codes. These product codes get inherited during bundling so that the functionality gets passed on to future generations of the AMI.
You should only be concerned with this particular message if your AMI is a paid AMI or somehow requires product-codes to be attached for it to function correctly.
Hope that helps.
|
|
Posts:
3
Registered:
4/27/08
|
|
|
|
Re: Server refused to allocate pty
Posted:
Apr 30, 2008 12:34 AM PDT
in response to: Marcin@AWS
|
|
|
I'm seeing the same issue as this. I get that I should've used the -fstab etc/fstab bundlevol flag when I bundled, but now that I've bundled and uploaded to S3 already (without having known to use the flag), does that mean my saved image isn't usable anymore, and I need to create a new AMI from scratch?
Thx,
Rajesh
|
|
Posts:
22
Registered:
12/11/06
|
|
|
|
Re: Server refused to allocate pty
Posted:
May 19, 2008 7:20 AM PDT
in response to: Rajesh Viswanat...
|
|
|
Just ran into this too. Quite annoying! for some reason it didn't have an effect until the first reboot of a newly-launched instance bundled from an FC8 instance -- which took a while to occur.
I tried rebuilding the fstab to look like what Marcin described on
Mar 28, 2008 2:45 PM GMT, then rebuilt an AMI on that node. A new node started with that AMI worked for the first boot, then failed again on the second boot.
On investigation, after the first boot, it appears the fstab had been reset again -- does the installation of the "ec2-ami-tools" RPM, which runs at every boot, replace it?
That's my guess.
I added this file (mode 755) as /etc/rc.d/rc4.d/S99devpts to work around it. It'd be nice to see it fixed "upstream" though.
|
|
Posts:
22
Registered:
12/11/06
|
|
|
|
Re: Server refused to allocate pty
Posted:
May 19, 2008 7:21 AM PDT
in response to: Justin Mason
|
|
|
|
|
here's the file....
|
|
Posts:
22
Registered:
12/11/06
|
|
|
|
Re: Server refused to allocate pty
Posted:
May 19, 2008 7:23 AM PDT
in response to: Rajesh Viswanat...
|
|
|
oh yeah, and Rajesh -- you can safely build a new AMI from a "broken" instance, as long as you fix your fstab first and add something like that workaround init script.
|
|
Posts:
22
Registered:
12/11/06
|
|
|
|
Re: Server refused to allocate pty
Posted:
May 19, 2008 8:23 AM PDT
in response to: Justin Mason
|
|
|
Ugh. Note that the file needs to be saved as "S99devpts", with a capital S -- the forum software here helpfully lower-cased it. :(
|
|
Posts:
162
Registered:
11/2/06
|
|
|
Posts:
22
Registered:
12/11/06
|
|
|
|
Re: Server refused to allocate pty
Posted:
Jun 11, 2008 1:59 AM PDT
in response to: Attila@AWS
|
|
|
great -- thanks Attila!
|
|
Posts:
20
Registered:
12/5/07
|
|
|
|
Re: Server refused to allocate pty
Posted:
Jul 17, 2008 10:35 AM PDT
in response to: Justin Mason
|
|
|
I am using ami-2b5fba42 (Amazon's Fedora 8 v1.07) and building instances off it still do not work properly. Should I reinstall the AMI Tools - is that the problem? If so, how do I remove the original (sorry, linux newbie question, I know)? I'm going to attempt to replace /etc/fstab with what was posted earlier, but just wanted to document that this problem seems to still be active.
|
|
Posts:
20
Registered:
12/5/07
|
|
|
|
Re: Server refused to allocate pty
Posted:
Jul 17, 2008 1:12 PM PDT
in response to: joehung
|
|
|
Just thought I'd give an update, I deleted everything in /usr/local/bin except rsync and installed an up to date copy of the ec2 ami tools, and everything seems to be working swimmingly. That said, someone at Amazon really ought to implement this fix on their end, so that other people no longer have this problem.
|
|
|
|
