CentOS 5.2 / x86_64 / download+build of rsync 3.0.4. Transferring data between two NetApp servers-- both NFS mounted on this (Linux) system. We're transferring from one directory tree to another with a script that looks like: for i in /data2/* /data3/* do /usr/local/bin/rsync -a --delete ${i}/ /data1/`basename $i` done For some files/dirs, rsync is failing: rsync: rename "/data1/bordner/Pfam/version_22.0/fasta/.PF09720.fa.8z0Imb" -> "Pfam/version_22.0/fasta/PF09720.fa": File too large (27) rsync: rename "/data1/bordner/Pfam/version_22.0/fasta/.PF09721.fa.0OoTTG" -> "Pfam/version_22.0/fasta/PF09721.fa": File too large (27) rsync: rename "/data1/bordner/Pfam/version_22.0/fasta/.PF09722.fa.83Sssc" -> "Pfam/version_22.0/fasta/PF09722.fa": File too large (27) rsync: rename "/data1/bordner/Pfam/version_22.0/fasta/.PF09723.fa.0ynM2H" -> "Pfam/version_22.0/fasta/PF09723.fa": File too large (27) rsync error: some files/attrs were not transferred (see previous errors) (code 23) at main.c(1040) [sender=3.0.4] The individual files are NOT "too large": [root@rcfcluster-rsync fasta]# cd /data3/bordner/Pfam/version_22.0/fasta/ [root@rcfcluster-rsync fasta]# ls -l PF09723.fa -rw-rw-r-- 1 m049922 bordner 21279 May 22 2008 PF09723.fa BUT, the directory is very large: [root@rcfcluster-rsync fasta]# ls -1 | wc -l 191811
The errors are almost certainly the fault of the destination filesystem, not of rsync. (Indeed, it's conceivable that the directory is hitting some NetApp-specific size limit, and the "File too large" message refers to that.) Try the copy with another tool, such as GNU cp, and see if you get similar errors.
Thanks for the quick reply. I'll try your suggestion and do a: [root@rcfcluster-rsync ~]# cp -a /data3/bordner/* /data1/bordner/ and see what that does. According to the NetApp, I'm nowhere near max files on that volume (but it certainly could be their bug...) I'll post note when "cp" finishes.
Spent some time chasing Red Herrings, but I think you are correct-- the dest file system is returning goofy errors... The file system has plenty of free space + inodes, but the "cp" of a small (135 byte) file fails with EFBIG (on the OPEN, no less). I'll chase this with NetApp....