zfs get all command hung
Gary Palmer
gpalmer at freebsd.org
Tue May 17 21:17:17 UTC 2011
On Tue, May 17, 2011 at 09:54:12PM +0100, a.smith at ukgrid.net wrote:
> Quoting Andriy Gapon <avg at FreeBSD.org>:
> >
> >procstat -kk <pid>
> >
>
> # procstat -kk 37975
> PID TID COMM TDNAME KSTACK
> 37975 100669 zfs - mi_switch+0x176
> sleepq_catch_signals+0x29e sleepq_wait_sig+0x16 _sleep+0x269
> clnt_vc_create+0x153 clnt_reconnect_call+0x64d nfs_request+0x215
> nfs_statfs+0x194 __vfs_statfs+0x28 kern_getfsstat+0x3fc
> syscallenter+0x1e5 syscall+0x4b Xfast_syscall+0xe2
>
> And actually I was thinking, as the all zfs get commands are hanging,
> I can run others and truss them of course. Here is the tail of a truss:
>
>
> NAME PROPERTY VALUE SOURCE
> write(1,"NAME PROPERTY VALU"...,58) = 58 (0x3a)
> mx1 type filesystem -
> write(1,"mx1 type file"...,53) = 53 (0x35)
> mx1 creation Mon Jan 17 12:08 2011 -
> write(1,"mx1 creation Mon "...,53) = 53 (0x35)
> mx1 used 78.2G -
> write(1,"mx1 used 78.2"...,53) = 53 (0x35)
> mx1 available 195G -
> write(1,"mx1 available 195G"...,53) = 53 (0x35)
> mx1 referenced 22K -
> write(1,"mx1 referenced 22K "...,53) = 53 (0x35)
> mx1 compressratio 1.27x -
> write(1,"mx1 compressratio 1.27"...,53) = 53 (0x35)
> fstat(4,{ mode=crw-rw-rw- ,inode=32,size=0,blksize=4096 }) = 0 (0x0)
> ioctl(4,TIOCGETA,0xffffc8c0) ERR#19 'Operation not
> supported by device'
> lseek(4,0x0,SEEK_SET) = 0 (0x0)
> lseek(4,0x0,SEEK_CUR) = 0 (0x0)
> getfsstat(0x0,0x0,0x1,0x0,0x80,0xa008) = 443 (0x1bb)
I'm no expert, but it looks more like you have a NFS filesystem mounted
on the system and for some reason system calls to list the mounted
filesystems are hanging due to the NFS mount. Is there a NFS filesystem
mounted on that box and is the NFS server available and responding to
NFS requests?
Gary
More information about the freebsd-fs
mailing list