[Gumstix-users] Server refused to allocate PTY (2.6.15 putty connection over BT)

[t***@bellsouth.net]

My bluetooth problems were a result of a configuration problem on my laptop. I'd forgotten that I had to configure the bluetooth network access (under "Select the type of service to offer remote devices") to "Allow other devices to access the internet/Lan via this computer".

The default is "Allow other devices to to create a private network with this adapter" and this sounds like it should work. But, after calling Kensington technical support (BT dongle vendor) on the first laptop I set up months ago, I was told to configure for "Allow other devices to access the internet/Lan via this computer". I'd forgotten to do this with my latest laptop configuration.

So, now I get a bluetooth connection can ping the gum from the laptop (and vice versa) .But, when I try to connect via putty, I get "Server refused to allocate PTY". This time I checked to make sure that this was unique to my 2.6.15 build by reloading the old buildroot. Everything works normal on the older buildroot

Back on 2.6.15, on the gum in /tmp/messages I see (after seeing "Server refused to allocate PTY" on the laptop):

Dec 31 16:14:54 gumstix authpriv.notice dropbear[2850]: password auth succeeded for 'root' from 192.168.0.1:1312
Dec 31 16:14:54 gumstix authpriv.warn dropbear[2850]: pty_allocate: openpty: No such file or directory
Dec 31 16:14:54 gumstix authpriv.warn dropbear[2850]: no pty was allocated, couldn't execute
Dec 31 16:14:54 gumstix authpriv.info dropbear[2850]: exit after auth (root): Exited normally

So, I'm guessing I've forgotten some configuration step, or have the default TARGETS changed and I'm missing something?

Thanks,

Dave Thomas

[Terry Markovich]

I also had this problem with ssh. It seems to be a problem
with /dev/pts being mounted by fstab, then being overwritten by udev. I
found a few other references to this problem, mostly with new migrations
to udev. My sloppy solution is to create a script in /etc/init.d/ named
S10pts and which contains:

#!/bin/sh
# remount /dev/pts
mount -t devpts none /dev/pts

make sure to chmod it executable (755). Now /dev/pts will get remounted
and your pty's will get created. I'm not sure what the real solution is,
probably something in udev.....

Terry Markovich

Post by t***@bellsouth.net
My bluetooth problems were a result of a configuration problem
on my laptop. I'd forgotten that I had to configure the
bluetooth network access (under "Select the type of service
to offer remote devices") to "Allow other devices to
access the internet/Lan via this computer".
The default is "Allow other devices to to create a private
network with this adapter" and this sounds like it should work.
But, after calling Kensington technical support (BT dongle vendor)
on the first laptop I set up months ago, I was told to configure
for "Allow other devices to access the internet/Lan via this
computer". I'd forgotten to do this with my latest laptop configuration.
So, now I get a bluetooth connection can ping the gum from the laptop
(and vice versa) .But, when I try to connect via putty, I get "Server
refused to allocate PTY". This time I checked to make sure that
this was unique to my 2.6.15 build by reloading the old buildroot.
Everything works normal on the older buildroot
Back on 2.6.15, on the gum in /tmp/messages I see (after seeing
Dec 31 16:14:54 gumstix authpriv.notice dropbear[2850]: password
auth succeeded for 'root' from 192.168.0.1:1312
openpty: No such file or directory
Dec 31 16:14:54 gumstix authpriv.warn dropbear[2850]: no pty was
allocated, couldn't execute
Dec 31 16:14:54 gumstix authpriv.info dropbear[2850]: exit after
auth (root): Exited normally
So, I'm guessing I've forgotten some configuration step, or have
the default TARGETS changed and I'm missing something?
Thanks,
Dave Thomas

[Dave Hylands]

Hi Terry,

Post by Terry Markovich
I also had this problem with ssh. It seems to be a problem
with /dev/pts being mounted by fstab, then being overwritten by udev. I
found a few other references to this problem, mostly with new migrations
to udev. My sloppy solution is to create a script in /etc/init.d/ named
#!/bin/sh
# remount /dev/pts
mount -t devpts none /dev/pts

Interesting, that fixes the ssh problem as well. After boot I see:

# mount
/dev/mtdblock1 on / type jffs2 (rw,noatime)
/proc on /proc type proc (rw,nodiratime)
sysfs on /sys type sysfs (rw)
devpts on /dev/pts type devpts (rw)
tmpfs on /tmp type tmpfs (rw)
udev on /dev type ramfs (rw)
/dev/mmcblk0p1 on /mnt/mmc type vfat
(rw,nodiratime,fmask=0022,dmask=0022,codepage=cp437,iocharset=iso8859-1)

If I try 'umount /dev/pts' it fails.
If I do 'mount /dev/pts' followed by mount I see /dev/pts showing up a
second time:

# mount
/dev/mtdblock1 on / type jffs2 (rw,noatime)
/proc on /proc type proc (rw,nodiratime)
sysfs on /sys type sysfs (rw)
devpts on /dev/pts type devpts (rw)
tmpfs on /tmp type tmpfs (rw)
udev on /dev type ramfs (rw)
/dev/mmcblk0p1 on /mnt/mmc type vfat
(rw,nodiratime,fmask=0022,dmask=0022,codepage=cp437,iocharset=iso8859-1)
devpts on /dev/pts type devpts (rw)

and ssh works fine.

--
Dave Hylands
Vancouver, BC, Canada
http://www.DaveHylands.com/

[Craig Hughes]

Post by Dave Hylands
Hi Terry,

Post by Terry Markovich
I also had this problem with ssh. It seems to be a problem
with /dev/pts being mounted by fstab, then being overwritten by udev. I
found a few other references to this problem, mostly with new
migrations
to udev. My sloppy solution is to create a script in /etc/init.d/ named
#!/bin/sh
# remount /dev/pts
mount -t devpts none /dev/pts

# mount
/dev/mtdblock1 on / type jffs2 (rw,noatime)
/proc on /proc type proc (rw,nodiratime)
sysfs on /sys type sysfs (rw)
devpts on /dev/pts type devpts (rw)
tmpfs on /tmp type tmpfs (rw)
udev on /dev type ramfs (rw)
/dev/mmcblk0p1 on /mnt/mmc type vfat
(rw,nodiratime,fmask=0022,dmask=0022,codepage=cp437,iocharset=iso8859-
1)
If I try 'umount /dev/pts' it fails.
If I do 'mount /dev/pts' followed by mount I see /dev/pts showing up a
# mount
/dev/mtdblock1 on / type jffs2 (rw,noatime)
/proc on /proc type proc (rw,nodiratime)
sysfs on /sys type sysfs (rw)
devpts on /dev/pts type devpts (rw)
tmpfs on /tmp type tmpfs (rw)
udev on /dev type ramfs (rw)
/dev/mmcblk0p1 on /mnt/mmc type vfat
(rw,nodiratime,fmask=0022,dmask=0022,codepage=cp437,iocharset=iso8859-
1)
devpts on /dev/pts type devpts (rw)
and ssh works fine.

Yup -- I just checked in a one-liner patch to S00udev which just adds
a "mount /dev/pts" after it creates /dev/pts in the newly mounts /dev

C

[Shai Rubin]

Yup -- I just checked in a one-liner patch to S00udev which just adds a
"mount /dev/pts" after it creates /dev/pts in the newly mounts /dev

Re: telnet server
2006-02-03 15:58
sshd is already on there (which is almost always preferable to >
telnetd).

Well, this is exactly my problem: sshd does not work (see log below).
This happened after I flushed my file system with a new
root_fs_arm_nofpu. I can use the console but not sshd. I wonder if I
used the right bootargs in u-boot. I'm not sure which kernel version I
have. I think it is linux-2.6.15gum because this is what I find in the
build_arm_nofpu directory. If it is "15gum", I wonder whether I need to
use the same bootargs specified for "13gum" in the tutorial.

If yes, how do I fix it? do I need to start the building process all
over again? that it, delete my "gumstix-buildroot"?

Shai

[Craig Hughes]

You can actually fix it on a "live" gumstix pretty easily -- just add
the line

mount /dev/pts

in /etc/init.d/S00udev

just after the line:

mkdir $udev_root/pts $udev_root/shm


C

Post by Craig Hughes
Yup -- I just checked in a one-liner patch to S00udev which just
adds a "mount /dev/pts" after it creates /dev/pts in the newly
mounts /dev

Re: telnet server
2006-02-03 15:58
sshd is already on there (which is almost always preferable to >
telnetd).

Well, this is exactly my problem: sshd does not work (see log below).
This happened after I flushed my file system with a new
root_fs_arm_nofpu. I can use the console but not sshd. I wonder if I
used the right bootargs in u-boot. I'm not sure which kernel

version I

Post by Craig Hughes
have. I think it is linux-2.6.15gum because this is what I find in

the

Post by Craig Hughes
build_arm_nofpu directory. If it is "15gum", I wonder whether I

need to

Post by Craig Hughes
use the same bootargs specified for "13gum" in the tutorial.

If yes, how do I fix it? do I need to start the building process
all over again? that it, delete my "gumstix-buildroot"?
Shai
-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through
log files
for problems? Stop! Download the new AJAX search engine that makes
searching your log files as easy as surfing the web. DOWNLOAD
SPLUNK!
http://sel.as-us.falkag.net/sel?
cmd=lnk&kid=103432&bid=230486&dat=121642
_______________________________________________
gumstix-users mailing list
https://lists.sourceforge.net/lists/listinfo/gumstix-users

[Alexandre Pereira Nunes]

Post by Craig Hughes
You can actually fix it on a "live" gumstix pretty easily -- just add
the line
mount /dev/pts
in /etc/init.d/S00udev
mkdir $udev_root/pts $udev_root/shm
C

I also put noauto option for /dev/pts entry at /etc/fstab, so that it
doesn't get mounted twice; it worked just fine.

[Craig Hughes]

Post by Alexandre Pereira Nunes

Post by Craig Hughes
You can actually fix it on a "live" gumstix pretty easily -- just add
the line
mount /dev/pts
in /etc/init.d/S00udev
mkdir $udev_root/pts $udev_root/shm
C

I also put noauto option for /dev/pts entry at /etc/fstab, so that it
doesn't get mounted twice; it worked just fine.

Good idea -- i'll add that.

C

[Craig Hughes]

Post by t***@bellsouth.net
My bluetooth problems were a result of a configuration problem on
my laptop. I'd forgotten that I had to configure the bluetooth
network access (under "Select the type of service to offer remote
devices") to "Allow other devices to access the internet/Lan via
this computer".
The default is "Allow other devices to to create a private network
with this adapter" and this sounds like it should work. But, after
calling Kensington technical support (BT dongle vendor) on the
first laptop I set up months ago, I was told to configure for
"Allow other devices to access the internet/Lan via this
computer". I'd forgotten to do this with my latest laptop
configuration.
So, now I get a bluetooth connection can ping the gum from the
laptop (and vice versa) .But, when I try to connect via putty, I
get "Server refused to allocate PTY". This time I checked to make
sure that this was unique to my 2.6.15 build by reloading the old
buildroot. Everything works normal on the older buildroot
Back on 2.6.15, on the gum in /tmp/messages I see (after seeing
Dec 31 16:14:54 gumstix authpriv.notice dropbear[2850]: password
auth succeeded for 'root' from 192.168.0.1:1312
openpty: No such file or directory
Dec 31 16:14:54 gumstix authpriv.warn dropbear[2850]: no pty was
allocated, couldn't execute
Dec 31 16:14:54 gumstix authpriv.info dropbear[2850]: exit after
auth (root): Exited normally
So, I'm guessing I've forgotten some configuration step, or have
the default TARGETS changed and I'm missing something?

That's the same problem others have reported. I suspect it's
something I messed up when moving to udev. I'll fix it as soon as I
figure out the problem.

C

[Craig Hughes]

Ok, figured out the problem in about 10 seconds. It looks like /dev/
pts is getting mounted *before* /dev is being mounted. So the /dev
mount is masking /dev/pts... duh! Will be fixed in a jiffy in the
buildroot.

C

Post by t***@bellsouth.net
My bluetooth problems were a result of a configuration problem on
my laptop. I'd forgotten that I had to configure the bluetooth
network access (under "Select the type of service to offer remote
devices") to "Allow other devices to access the internet/Lan via
this computer".
The default is "Allow other devices to to create a private network
with this adapter" and this sounds like it should work. But, after
calling Kensington technical support (BT dongle vendor) on the
first laptop I set up months ago, I was told to configure for
"Allow other devices to access the internet/Lan via this
computer". I'd forgotten to do this with my latest laptop
configuration.
So, now I get a bluetooth connection can ping the gum from the
laptop (and vice versa) .But, when I try to connect via putty, I
get "Server refused to allocate PTY". This time I checked to make
sure that this was unique to my 2.6.15 build by reloading the old
buildroot. Everything works normal on the older buildroot
Back on 2.6.15, on the gum in /tmp/messages I see (after seeing
Dec 31 16:14:54 gumstix authpriv.notice dropbear[2850]: password
auth succeeded for 'root' from 192.168.0.1:1312
openpty: No such file or directory
Dec 31 16:14:54 gumstix authpriv.warn dropbear[2850]: no pty was
allocated, couldn't execute
Dec 31 16:14:54 gumstix authpriv.info dropbear[2850]: exit after
auth (root): Exited normally
So, I'm guessing I've forgotten some configuration step, or have
the default TARGETS changed and I'm missing something?
Thanks,
Dave Thomas
-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through
log files
for problems? Stop! Download the new AJAX search engine that makes
searching your log files as easy as surfing the web. DOWNLOAD
SPLUNK!
http://sel.as-us.falkag.net/sel?
cmd=lnk&kid=103432&bid=230486&dat=121642
_______________________________________________
gumstix-users mailing list
https://lists.sourceforge.net/lists/listinfo/gumstix-users

[Shai Rubin]

A general question, does the 2.6.15 stable? If not, how can I obtain the
most recent stable release for the gumstix linux kernel.


Shai

Post by Craig Hughes
Ok, figured out the problem in about 10 seconds. It looks like /dev/
pts is getting mounted *before* /dev is being mounted. So the /dev
mount is masking /dev/pts... duh! Will be fixed in a jiffy in the
buildroot.
C

My bluetooth problems were a result of a configuration problem on my
laptop. I'd forgotten that I had to configure the bluetooth network
access (under "Select the type of service to offer remote devices")
to "Allow other devices to access the internet/Lan via this computer".
The default is "Allow other devices to to create a private network
with this adapter" and this sounds like it should work. But, after
calling Kensington technical support (BT dongle vendor) on the first
laptop I set up months ago, I was told to configure for "Allow other
devices to access the internet/Lan via this computer". I'd forgotten
to do this with my latest laptop configuration.
So, now I get a bluetooth connection can ping the gum from the laptop
(and vice versa) .But, when I try to connect via putty, I get "Server
refused to allocate PTY". This time I checked to make sure that
this was unique to my 2.6.15 build by reloading the old buildroot.
Everything works normal on the older buildroot
Back on 2.6.15, on the gum in /tmp/messages I see (after seeing
Dec 31 16:14:54 gumstix authpriv.notice dropbear[2850]: password auth
succeeded for 'root' from 192.168.0.1:1312
openpty: No such file or directory
Dec 31 16:14:54 gumstix authpriv.warn dropbear[2850]: no pty was
allocated, couldn't execute
Dec 31 16:14:54 gumstix authpriv.info dropbear[2850]: exit after auth
(root): Exited normally
So, I'm guessing I've forgotten some configuration step, or have the
default TARGETS changed and I'm missing something?
Thanks,
Dave Thomas
-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through
log files
for problems? Stop! Download the new AJAX search engine that makes
searching your log files as easy as surfing the web. DOWNLOAD SPLUNK!
http://sel.as-us.falkag.net/sel? cmd=lnk&kid=103432&bid=230486&dat=121642
_______________________________________________
gumstix-users mailing list
https://lists.sourceforge.net/lists/listinfo/gumstix-users

-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log
files
for problems? Stop! Download the new AJAX search engine that makes
searching your log files as easy as surfing the web. DOWNLOAD SPLUNK!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=103432&bid=230486&dat=121642
_______________________________________________
gumstix-users mailing list
https://lists.sourceforge.net/lists/listinfo/gumstix-users

[Dave Hylands]

Hi Shai,

Post by Shai Rubin
A general question, does the 2.6.15 stable? If not, how can I obtain the
most recent stable release for the gumstix linux kernel.

I would say that 2.6.15 should not be considered stable yet. It's
getting closer every day as people start to use it and report problems
that they encounter.

The latest stable release can be found on SourceForge:
http://sourceforge.net/project/showfiles.php?group_id=98784

It's currently at SVN release 773. Version 775 is where the 2.6.15
stuff started to happen (so 773 is the last version that uses 2.6.11).

--
Dave Hylands
Vancouver, BC, Canada
http://www.DaveHylands.com/

[Craig Hughes]

The most recent stable release is always the binary on sf.net/
projects/gumstix -- those binaries are labelled with the subversion
trunk revision number to which they correspond, so you can use that
revision from subversion to rebuild them. The HEAD of trunk is
generally "quite stable" -- that is, it's more or less a release
candidate. If there's a big change underway which is likely to break
the build somewhat, I'll do that on a branch first, then merge back
to trunk once it's in "release candidate" stage. The switch to
2.6.15 from 2.6.11 introduced quite a few major changes, which is why
there are "issues" with the trunk right now. 2.6.15 was good enough
on the branch though that I felt it was time to merge it back in, to
get enough people working with it to find all these remaining wrinkles.

C

Post by Shai Rubin
A general question, does the 2.6.15 stable? If not, how can I
obtain the most recent stable release for the gumstix linux kernel.
Shai

Post by Craig Hughes
Ok, figured out the problem in about 10 seconds. It looks like /
dev/ pts is getting mounted *before* /dev is being mounted. So
the /dev mount is masking /dev/pts... duh! Will be fixed in a
jiffy in the buildroot.
C

Post by t***@bellsouth.net
My bluetooth problems were a result of a configuration problem
on my laptop. I'd forgotten that I had to configure the
bluetooth network access (under "Select the type of service to
offer remote devices") to "Allow other devices to access the
internet/Lan via this computer".
The default is "Allow other devices to to create a private
network with this adapter" and this sounds like it should work.
But, after calling Kensington technical support (BT dongle
vendor) on the first laptop I set up months ago, I was told to
configure for "Allow other devices to access the internet/Lan
via this computer". I'd forgotten to do this with my latest
laptop configuration.
So, now I get a bluetooth connection can ping the gum from the
laptop (and vice versa) .But, when I try to connect via putty, I
get "Server refused to allocate PTY". This time I checked to
make sure that this was unique to my 2.6.15 build by reloading
the old buildroot. Everything works normal on the older buildroot
Back on 2.6.15, on the gum in /tmp/messages I see (after seeing
Dec 31 16:14:54 gumstix authpriv.notice dropbear[2850]: password
auth succeeded for 'root' from 192.168.0.1:1312
pty_allocate: openpty: No such file or directory
Dec 31 16:14:54 gumstix authpriv.warn dropbear[2850]: no pty was
allocated, couldn't execute
Dec 31 16:14:54 gumstix authpriv.info dropbear[2850]: exit after
auth (root): Exited normally
So, I'm guessing I've forgotten some configuration step, or have
the default TARGETS changed and I'm missing something?
Thanks,
Dave Thomas
-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep
through log files
for problems? Stop! Download the new AJAX search engine that makes
searching your log files as easy as surfing the web. DOWNLOAD SPLUNK!
http://sel.as-us.falkag.net/sel?
cmd=lnk&kid=103432&bid=230486&dat=121642
_______________________________________________
gumstix-users mailing list
https://lists.sourceforge.net/lists/listinfo/gumstix-users

-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through
log files
for problems? Stop! Download the new AJAX search engine that makes
searching your log files as easy as surfing the web. DOWNLOAD SPLUNK!
http://sel.as-us.falkag.net/sel?
cmd=lnk&kid=103432&bid=230486&dat=121642
_______________________________________________
gumstix-users mailing list
https://lists.sourceforge.net/lists/listinfo/gumstix-users

-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through
log files
for problems? Stop! Download the new AJAX search engine that makes
searching your log files as easy as surfing the web. DOWNLOAD SPLUNK!
http://sel.as-us.falkag.net/sel?
cmd=lnk&kid=103432&bid=230486&dat=121642
_______________________________________________
gumstix-users mailing list
https://lists.sourceforge.net/lists/listinfo/gumstix-users

[t***@bellsouth.net]

Yep, seems to work fine after adding the mount /dev/pts in the S00udev file. Thanks!

Maybe tomorrow I can finally try out the UCB1400 touch.

Thanks!

Dave Thomas

[Adam McLeod]

I checked out the latest build, and subversion reported no change to
S00udev, but it did change something in init.d that looked a whole lot
like the contents of S00udev. What is going on there?

Adam

Post by t***@bellsouth.net
Yep, seems to work fine after adding the mount /dev/pts in the S00udev file. Thanks!

[Dave Hylands]

Hi Adam,

Post by Adam McLeod
I checked out the latest build, and subversion reported no change to
S00udev, but it did change something in init.d that looked a whole lot
like the contents of S00udev. What is going on there?

My guess is that S00udev is generated from the init-udev file that was
modified. This is further confirmed by the fact that there is no
S00udev in the target_skeleton.

--
Dave Hylands
Vancouver, BC, Canada
http://www.DaveHylands.com/

[Craig Hughes]

Post by Dave Hylands
Hi Adam,

Post by Adam McLeod
I checked out the latest build, and subversion reported no change to
S00udev, but it did change something in init.d that looked a whole lot
like the contents of S00udev. What is going on there?

My guess is that S00udev is generated from the init-udev file that was
modified. This is further confirmed by the fact that there is no
S00udev in the target_skeleton.

Yup -- init-udev gets copied to S00udev; it's not coming from the
target_skeleton because some people might not install the udev
package, and target_skeleton gets on there no matter what packages
you've selected.

C