Gluset to Ganesha (3 hosts and HA Cluster) Confirm bug present in 3.7.13 and 3.7.14 (candidate) centos builds Reproducd in all hosts. [root@xintel2 ~]# LD_PRELOAD=/usr/lib64/libasan.so.2 /usr/bin/ganesha.nfsd -F -f /etc/ganesha/ganesha.conf -N NIV_EVENT *** Error in `/usr/bin/ganesha.nfsd': free(): invalid size: 0x00006060000083c0 *** ======= Backtrace: ========= /lib64/libc.so.6(+0x7d053)[0x7f30834d8053] /lib64/libglusterfs.so.0(_gf_msg+0x347)[0x7f307e9f7747] /usr/lib64/glusterfs/3.7.14/xlator/mount/api.so(notify+0x18a)[0x7f3072724fda] /lib64/libglusterfs.so.0(xlator_notify+0x36)[0x7f307e9f3eb6] /lib64/libglusterfs.so.0(glusterfs_graph_activate+0xac)[0x7f307ea438bc] /lib64/libgfapi.so.0(+0x8ded)[0x7f307ecb4ded] /lib64/libgfapi.so.0(+0x8f99)[0x7f307ecb4f99] /lib64/libgfrpc.so.0(rpc_clnt_handle_reply+0x90)[0x7f307e7c4620] /lib64/libgfrpc.so.0(rpc_clnt_notify+0x1df)[0x7f307e7c48ff] /lib64/libgfrpc.so.0(rpc_transport_notify+0x23)[0x7f307e7c07c3] /usr/lib64/glusterfs/3.7.14/rpc-transport/socket.so(+0x6644)[0x7f30714e1644] /usr/lib64/glusterfs/3.7.14/rpc-transport/socket.so(+0x9284)[0x7f30714e4284] /lib64/libglusterfs.so.0(+0x9537a)[0x7f307ea6537a] /lib64/libpthread.so.0(+0x7dc5)[0x7f3083e83dc5] /lib64/libc.so.6(clone+0x6d)[0x7f3083551ced] ======= Memory map: ======== 7fff7000-8fff7000 rw-p 00000000 00:00 0 8fff7000-2008fff7000 ---p 00000000 00:00 0 2008fff7000-10007fff8000 rw-p 00000000 00:00 0 600000000000-602000000000 ---p 00000000 00:00 0 602000000000-602000010000 rw-p 00000000 00:00 0
AFAIK nobody has made a habit of running ASAN on our code. There are dozens of such tools out there. Each finds different things, and each yields false positives. We've run many of them, but do not have the resources to run every single one and sort through the results. If somebody wants to volunteer, that would be awesome.
There is an issue with upcall processing in ganesha. Could you please turn off that and re-test? gluster v set <volname> features.cache-invalidation off.
This bug is getting closed because GlusteFS-3.7 has reached its end-of-life. Note: This bug is being closed using a script. No verification has been performed to check if it still exists on newer releases of GlusterFS. If this bug still exists in newer GlusterFS releases, please reopen this bug against the newer release.