Hi Matt, > What is still unresolved is the seqfault of the "cp" command when I try > to copy > something on my pvfs2 mounted tree: > -----------8<---------------- > [EMAIL PROTECTED] pvfs2]# mount | grep pvfs2 > tcp://node1:3334/pvfs2-fs on /mnt/pvfs2 type pvfs2 (rw) > [EMAIL PROTECTED] pvfs2]# pwd > /mnt/pvfs2 > [EMAIL PROTECTED] pvfs2]# ls -l > total 6436 > drwxrwxrwx 1 root root 4096 Jun 6 22:29 lost+found/ > -rw-rw-rw- 1 root root 3291038 Jun 7 09:26 testfile > [EMAIL PROTECTED] pvfs2]# cp testfile testfile3 > Segmentation fault > [EMAIL PROTECTED] pvfs2]# cp testfile /root/testfile2 > cp: setting attributes for UH?H?: Operation not supported
Umm.. this is weird.. I will take a look at this on your setup.. > The last message indicates some attribute setting problems. > Interestingly, "scp" fails > like cp when copying out of the pvfs2 filesystem while it succeeds > without seqfault > if the target is remains on pvfs2: Hmm.. it looks like some unaligned accesses may have crept into CVS head causing cp/other programs to crash when using pvfs2..? Pete, did you noticed anything that went in recently that may cause unaligned access violations? BTW: Are there any tools for detecting this sort of thing? does valgrind on opteron have an unaligned memory access plugin or something? I will take a look at it and see if I can reproduce it... Thanks, Murali > > -----------8<---------------- > [EMAIL PROTECTED] pvfs2]# scp testfile /root/testfile4 > cp: setting attributes for UH?H?: Operation not supported > [EMAIL PROTECTED] pvfs2]# scp testfile testfile4 > ----------->8---------------- > > > My storage areas are on ext3 filesystems. My pvfs2-fs.conf and pvfs2tab > are attached. > I'm running yesterday's (06/06/06) CVS version. > > ... Matt > > _______________________________________________ Pvfs2-users mailing list [email protected] http://www.beowulf-underground.org/mailman/listinfo/pvfs2-users
