xrootdfs crash

[dougbenjamin]

Hi,

We are seeing this error in the system logs

May 1 14:54:48 atl008 kernel: xrootdfs[3249] trap divide error rip:336920819a rsp:2af4fdc67710 error:0

for a static mount xrootdfs mount point -
xrootdfs /atlfs03/atlas fuse rdr=root://atlfs03.phy.duke.edu:1094//atlas,uid=54657 0 0

[root@atl008 ~]# yum list installed xrootd-fuse
Loaded plugins: downloadonly, kernel-module
Excluding Packages from Extra Packages for Enterprise Linux 5 - x86_64
Finished
Excluding Packages from VDT RPM repository - development versions for Redhat Enterprise Linux 5 and compatible
Finished
Installed Packages
xrootd-fuse.x86_64 1:3.3.1-1.slc5.xu installed

Here is the OS details -

[benjamin@atl008 ~]$ uname -a
Linux atl008.phy.duke.edu 2.6.18-348.4.1.el5 #1 SMP Tue Apr 16 15:42:58 EDT 2013 x86_64 x86_64 x86_64 GNU/Linux
[benjamin@atl008 ~]$ cat /etc/redhat-release
Scientific Linux release 5.9 (Boron)

Any ideas how I can get around this issue.

Thanks,

Doug

[xrootd-dev]

Hi Doug,

I don't have an idea at this point but have two questions: is there a core file left behind? how often does this happen?

regards,
Wei Yang | yangw@slac.stanford.edu | 650-926-3338(O)

[dougbenjamin]

Hi,

More often than I would like. It appears to happen regularly. Not
sure how I can capture the core dump, when
the mount is stated from /etc/fstab (or autofs for that mater). Any
clues to get the core would be helpful

Thanks,

Doug

On 05/01/2013 02:21 PM, xrootd-dev wrote:

Hi Doug,

I don't have an idea at this point but have two questions: is there a
core file left behind? how often does this happen?

regards,
Wei Yang | yangw@slac.stanford.edu | 650-926-3338(O)

—
Reply to this email directly or view it on GitHub
#11 (comment).

[xrootd-dev]

I supposed you use a script to start xrootdfs in autofs? (you can use the script in /etc/fstab as well). in the script, can you add

cd /tmp
ulimit -c unlimited

regards,
Wei Yang | yangw@slac.stanford.edu | 650-926-3338(O)

On May 1, 2013, at 12:37 PM, dougbenjamin wrote:

Hi,

More often than I would like. It appears to happen regularly. Not
sure how I can capture the core dump, when
the mount is stated from /etc/fstab (or autofs for that mater). Any
clues to get the core would be helpful

Thanks,

Doug

[dougbenjamin]

How does one use a script in /etc/fstab ? I ask because I see that we are having trouble getting the automounter to remount xrootdfs mounts after they have been unmounted when running under SL6 (???!!!???)

[wyang007]

xrootdfs /xrootd/atlas fuse rdr=root://atl-xrdr:11094//atlas/xrootd,uid=atldq2 1 2

(the first xrootdfs can also be a script)

Wei Yang | yangw@slac.stanford.edu | 650-926-3338(O)

On Sep 2, 2013, at 3:25 PM, dougbenjamin notifications@github.com wrote:

How does one use a script in /etc/fstab ? I ask because I see that we are having trouble getting the automounter to remount xrootdfs mounts after they have been unmounted when running under SL6 (???!!!???)

—
Reply to this email directly or view it on GitHub.

[abh3]

I don't think this has been reproduced as of late. Has it? I'll reopen this if this is still a problem.