emerge hangs

Message

javeree · Post by **javeree** » Fri Jan 25, 2013 10:44 am

I have one PC, where emerge immediately hangs (no output at all). I cannot even end it with Ctrl-C.

From another tty, I can run ps aux | grep emerge and get

root 19605 9.9 1.4 19892 14372 pts/2 D+ 11:03 0:01 /usr/bin/python3.2 /usr/bin/emerge -uDNatv @world
root 19609 0.0 0.0 4624 780 pts/1 S+ 11:04 0:00 grep --colour=auto emerge

This D+ seems to indicate the sw is in uninterruptable sleep waiting for something. However, tehre is no indication what is blocking the program.
I could try to run 'strace emerge', but I don't have strace on this PC and I obviously cannot emerge it either.
Any suggestions on handling this ?

javeree · Post by **javeree** » Fri Jan 25, 2013 10:45 am

Never mind, I realized that I mount portage over nfs, and it is this mount that hangs somehow.

RobPearce · Post by **RobPearce** » Mon Mar 25, 2013 8:39 am

I have (at least) two machines that have started suffering this hang some time this year. Neither of them have /usr/portage on NFS but they do have the PKGDIR on NFS, and they hang during the "Performing Global Updates:" step:

Code: Select all

ezra ~ # emerge -C =sys-kernel/vanilla-sources-2.6.38.7

Performing Global Updates:
(Could take a couple of minutes if you have a lot of binary packages.)
  .='update pass'  *='binary update'  #='/var/db update'  @='/var/db move'
  s='/var/db SLOT move'  %='binary move'  S='binary SLOT move'
  p='update /etc/portage/package.*'
/usr/portage/profiles/updates/2Q-2011............................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................
/usr/portage/profiles/updates/3Q-2011..........................
/usr/portage/profiles/updates/4Q-2011.............................................................................................................
/usr/portage/profiles/updates/1Q-2012.......................
/usr/portage/profiles/updates/2Q-2012........................
/usr/portage/profiles/updates/3Q-2012.....................................................
/usr/portage/profiles/updates/4Q-2012.............................................................
/usr/portage/profiles/updates/1Q-2013................................................................................................

Running ps shows a "D+" state for emerge. Why has this suddenly broken when it worked absolutely fine up until now?

Randy Andy · Post by **Randy Andy** » Mon Mar 25, 2013 10:20 am

Try to stay patient for up to 15 minutes to get the process finish.

I have had the same behavior some times.

On slow machines it could take too much time, especially if the world file is full of superfluous entries which came into, if you forget to use the --oneshot option.

Only if this fails, I would state you have a problem...

Much success, Andy.

RobPearce · Post by **RobPearce** » Mon Mar 25, 2013 6:48 pm

Randy Andy wrote:Try to stay patient for up to 15 minutes to get the process finish.

I have had the same behavior some times.

On slow machines it could take too much time, especially if the world file is full of superfluous entries which came into, if you forget to use the --oneshot option.

Only if this fails, I would state you have a problem...

Much success, Andy.

It's WWWAAAAAAYYYYYYY more than 15 minutes! I left it running OVERNIGHT and it still hadn't finished.

However, I've established that the lock-up can be prevented by mounting the PKGDIR with "nolock". Of course, this is exactly the wrong solution, because PKGDIR is one of those places where locking is important, especially if shared between machines.

Randy Andy · Post by **Randy Andy** » Mon Mar 25, 2013 8:41 pm

RobPearce,

for your NFS constellation this might be a different thing and is known to bring more points of failures into.

My comment refers more to the initial poster, so let's see what he state.