Description of problem: nfs mounts of solaris 8 disks larger than about 1G don't show correct size if mounted vers=3. vers=2 mount shows correct size, as does i686 vers=3. Disk of size 500G works fine. May be related to bugzilla 162259. Version-Release number of selected component (if applicable): nfs-utils-1.0.6-33EL, update 5 plus rhn to date. How reproducible: completely Steps to Reproduce: 1.mount -o vers=3 sol8:/bigdisk /local 2.strace df -k /local/. 3. Actual results: statfs("/disk/other", {f_type="NFS_SUPER_MAGIC", f_bsize=18446744073709551615, f_blocks=18446744073709551615, f_bfree=18446744073709551615, f_bavail=18446744073709551615, f_files=18446744073709551615, f_ffree=18446744073709551615, f_fsid={0, 0}, f_namelen=255, f_frsize=0}) = 0 Expected results: same command on i686 machine, nfs-utils-1.0.6-33EL: statfs64("/disk/other", 84, {f_type="NFS_SUPER_MAGIC", f_bsize=32768, f_blocks=121891376, f_bfree=94671452, f_bavail=94671452, f_files=4294967295, f_ffree=4294967295, f_fsid={0, 0}, f_namelen=255, f_frsize=0}) = 0 stat64("/disk/other", {st_mode=S_IFDIR|0755, st_size=4096, ...}) = 0 Additional info: same command after mount -o vers=2 on x86_64: statfs("/a", {f_type="NFS_SUPER_MAGIC", f_bsize=8192, f_blocks=487565504, f_bfree=378298504, f_bavail=378298504, f_files=0, f_ffree=0, f_fsid={0, 0}, f_namelen=255, f_frsize=0}) = 0 stat("/a", {st_mode=S_IFDIR|0755, st_size=4096, ...}) = 0
What kernel version are you using? If its the bug I'm thinking of this as already been fixed in the U4 or U5 release
I'm thinking that this is a dup of https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=153094
#1 installed update 5 with redhat network updates as of yesterday. rebooted to uname -a: Linux amd2.ig.utexas.edu 2.4.21-32.0.1.ELsmp #1 SMP Tue May 17 17:46:36 EDT 2005 x86_64 x86_64 x86_64 GNU/Linux #2 I searched for nfs v3 and found only 162252 before I submitted a bug report I can't access bug 153904 now with my bugzilla privileges. Maybe this is a "user new to bugzilla problem" - if so, apologies in advance, but I tried.
I'm not sure why they the permission so tight in that bz, so I put some nfs-utils in http://people.redhat.com/steved/rhel3/ which contain that patch from bz153094. Give it a try to see if fixes this problem....
sorry to be so long getting back to you. I guess I don't see where this is necessarily an nfsutils problem if the statfs system call returns bad data for v3 - I shouldn't have mentioned nfs-utils, but I didn't then know how little of the operational nfs code nfs-utils has in it.
This problem has been resolved, although I don't understand exactly how. Some further information: the solaris 8 server filesystem was a Sun SAMfs filesystem (version 4.2.8). When this server code was upgraded to 4.3.15, apparently the df problem fixed itself. This wasn't noticed for a few days, since there was no expectation that the server side had anything to do with it. It may also be that the server was rebooted and it had nothing at all to do with the SAMfs software upgrade. now, strace df -k shows: statfs("/disk/other", {f_type="NFS_SUPER_MAGIC", f_bsize=32768, f_blocks=121891328, f_bfree=45927232, f_bavail=45927232, f_files=17776096, f_ffree=17196352, f_fsid={0, 0}, f_namelen=255, f_frsize=0}) = 0