Ralus segfaults on RHEL4 x86-64

[Cott_Lang]

3204960 Tue Mar 13 03:42:08 2007 : FileIsNFS ustat(): stat=0 f_tfree=72905998 f_tinode=72905998 f_fname=
43204960 Tue Mar 13 03:42:08 2007 : FileIsNFS: Is false for 39716113995-14402826-20070208-140615.xml
45a08960 Tue Mar 13 03:42:08 2007 : Function vfmObjCtl, VxMS resolved the error code to this message: system call failed
45a08960 Tue Mar 13 03:42:08 2007 : vx_read_ext: vfmObjCtl: ret_val : 0XB
45a08960 Tue Mar 13 03:42:08 2007 : VX_LoadEAbuffer: vx_read_ext failed. err = -536870075
45a08960 Tue Mar 13 03:42:08 2007 : ====> VX_OpenObj returning 0
43204960 Tue Mar 13 03:42:08 2007 : FileIsNFS statvfs(): stat=0 vbuf.f_files=77856768 f_ffree=72905998 f_fsid=0
43204960 Tue Mar 13 03:42:08 2007 : VX_GetObjInfo: os info not complete: val 1
43204960 Tue Mar 13 03:42:08 2007 : ====> VX_OpenObj
43204960 Tue Mar 13 03:42:08 2007 : Versioned file dblk osinfo, os_spec_size=127
43204960 Tue Mar 13 03:42:08 2007 : vx_stat_by_name: Getting stat of 14402826-20070208-140615.xml
43204960 Tue Mar 13 03:42:08 2007 : OpenFileUnix: bec_flags=1, Mode : 20
Segmentation fault

Is there a recommended procedure for debugging this further? We have two systems that ralus dies on after backing up successfully one time. I'll enable core dumps to see what we can get ...

[Sandy_Sandifer]

Hi Cott,

It's possible that the filename including the path have exceeded 255 characters. Would you be able to try the following:
1. Check to see if you have any paths (including the filename) that you're backing up that exceed 255 characters in your selections list or on the linux box. Take that path and copy it into word and run a count on the characters just to be sure.
2. If you find some that do exceed that, try doing backups of those to see if the same issue exists. You could also try doing backups of all the others that don't exceed that limit and see if those always work.

If that doesn't work you can also try unchecking "follow local mount points" in the properties of the backup job and see if the same issue occurs.

Let me know how that goes, and yes, if you have the core dump available, post that (the part that applies to the time of failure)...which brings up a good point: make sure you're looking at the time of failure in the log and compare that to the times in the debug and core dump.

Thanks.