On 10/31/2017 09:39 AM, Shyam Ranganathan wrote:
On 10/27/2017 09:15 AM, Vaibhav Vaingankar wrote:
Hello,
We are building GlusterFS from source. The build and installation is
successful, but while running `run-tests.sh` script we have
encountered the following error
Result: FAIL
./tests/basic/0symbol-check.t: bad status 1
*********************************
* REGRESSION FAILED *
* Retrying failed tests in case *
* we got some spurous failures *
*********************************
Your problem seems to be due to glusterd-snapshot.c using lstat instead
of sys_lstat. This is fixed in master here
https://review.gluster.org/#/c/16820/
If that patch is applied, you should be able to progress beyond this
issue. So you are looking at a genuine issue not (yet) a problem with
the tests as such.
I am attempting to understand how this passed (and continues to pass) in
the 3.10 release branch.
There seems to be a bug in our test case that makes it pass in the
regression machines, but fails elsewhere (looks like we are looking for
an incorrect/newer/older symbol on the regression machines).
This bug is filed to track and fix the problem in the test case:
https://bugzilla.redhat.com/show_bug.cgi?id=1508025
Further, the mentioned patch above is now ported to the 3.10 branch as
well ( https://review.gluster.org/18602 ), and should address the
specific issue in a future release (3.10.8 possibly).
Shyam
_______________________________________________
Gluster-devel mailing list
Gluster-devel@gluster.org
http://lists.gluster.org/mailman/listinfo/gluster-devel
