Re: failure in NetBSD while running as root

[Jeff Rizzo <riz%netbsd.org@localhost>]

On 3/29/11 2:14 AM, Julio Merino wrote:
> On Mon, Mar 28, 2011 at 5:31 PM, Jeff Rizzo<riz%netbsd.org@localhost>  wrote:
> > All I can see at this point is that atf-run itself seems to quit at the end
> > of that test, even if I remove the cleanup (which seems to fail, because
> > things have already been unmounted).
> Unfortunately, I can't get much information from the messages you provided :-(
>
> Are you running very -current sources?  There was a bug-fix for
> atf-run recently, although I don't think it's related.  That said, the
> code in atf-run is utterly convoluted so I'm wouldn't be surprised to
> find other bugs in it or if the mentioned bugfix resolved your issue.

Yes, it's running very -current.  I use this machine to run tests of 
NetBSD as it's built on the release engineering cluster.

> Do you see the problems if you run atf-run by itself?  (I.e. no
> atf-report involved.)  A raw trace of that would be more useful.

Yes, I do, but the trace isn't much more helpful.   I just happened to 
get a failure in fs/psshfs this time:

g4:riz  /usr/tests/fs/psshfs> sudo atf-run
Password:
Content-Type: application/X-atf-tps; version="2"

info: atf.version, Automated Testing Framework 0.12 (atf-0.12)
info: tests.root, /usr/tests/fs/psshfs
info: time.start, Tue Mar 29 11:26:19 PDT 2011
info: uname.sysname, NetBSD
info: uname.nodename, g4
info: uname.release, 5.99.48
info: uname.version, NetBSD 5.99.48 (GENERIC) #0: Sun Mar 27 23:08:23 
UTC 2011 
builds%b8.netbsd.org@localhost:/home/builds/ab/HEAD/macppc/201103271700Z-obj/home/builds/ab/HEAD/src/sys/arch/macppc/compile/GENERIC
info: uname.machine, macppc
info: env, PWD=/usr/tests/fs/psshfs
info: env, LANG=en_US.ISO8859-15
info: env, MAIL=/var/mail/riz
info: env, HOME=/home/riz
info: env, SUDO_GID=100
info: env, LC_ALL=en_US.ISO8859-15
info: env, 
PATH=/home/riz/bin:/home/riz/bin.macppc:/usr/local/bin:/usr/pkg/bin:/usr/bin:/usr/local/sbin:/usr/pkg/sbin:/usr/sbin:/bin:/sbin:/usr/X11R7/bin:/usr/games
info: env, SUDO_UID=6004
info: env, USERNAME=root
info: env, TERM=xterm-color
info: env, USER=root
info: env, LC_CTYPE=de_DE.ISO8859-15
info: env, LOGNAME=root
info: env, SHELL=/bin/csh
info: env, SUDO_USER=riz
info: env, SUDO_COMMAND=/usr/bin/atf-run
tps-count: 1
tp-start: t_psshfs, 3
tc-start: inode_nos
tc-so:Setting up SSH server configuration
tc-so:Executing command [ cp /usr/libexec/sftp-server . ]
tc-so:Executing command [ cp /usr/tests/fs/psshfs/ssh_host_key . ]
tc-so:Executing command [ cp /usr/tests/fs/psshfs/ssh_host_key.pub . ]
tc-so:Executing command [ chmod 400 ssh_host_key ]
tc-so:Executing command [ chmod 444 ssh_host_key.pub ]
tc-so:SSH server started (pid 14187)
tc-so:Setting up SSH client configuration
tc-so:Executing command [ ssh-keygen -f ssh_user_key -t rsa -b 1024 -N  -q ]
tc-so:Executing command [ cp ssh_user_key.pub authorized_keys ]
tc-so:Executing command [ chmod 600 authorized_keys ]
tc-so:Executing command [ mount -t psshfs -o 
-F=/tmp/atf-run.14884b/ssh_config localhost:/tmp/atf-run.14884b/root mnt ]
tc-so:Executing command [ ./ne_inodes.sh root/dir root/dir/file1 ]
tc-so:Executing command [ ./ne_inodes.sh root/dir root/dir/file2 ]
tc-so:Executing command [ ./ne_inodes.sh root/dir/file1 root/dir/file2 ]
tc-so:Executing command [ ./ne_inodes.sh root/file3 root/file4 ]
tc-se:umount: mnt: not currently mounted
tc-so:Stopping SSH server (pid 14187)
tc-so:Server output was:
tc-end: inode_nos, passed
atf-run: ERROR: XXX: Cannot get information of /tmp/atf-run.14884b/mnt;
atf-run: lstat(2) failed: Device not configured
g4:riz  /usr/tests/fs/psshfs>


I have actually removed the umount, and while I don't get the error, it 
just fails sometimes.
> Also, what happens if you run t_pssfs by hand?  Does the test work or
> it fails?  (mkdir foo ; ../t_psshfs inode_nos ; ./t_psshfs
> inode_nos:cleanup)

The couple times I did it, it worked OK... but then again, the atf-run 
doesn't fail every time.  (only most of them.

+j