A.2.4 Error messages (MDS (sfcfsmg daemon))

This section explains the MDS (sfcfsmg daemon) error messages of the GFS Shared File System.

ERROR: sfcfs_mds:0001: v_count != 0 inode=ino_pointer ino=ino_num

Explanation

An error was detected in releasing the i-node (ino_num) in the file system.

Response

Collect a file output with FJQSS, fjsnap, or pclsnap for troubleshooting information then contact field engineers.

ERROR: sfcfs_mds:1001: MDS is failed

Explanation

MDS failed in recovery mounting.

Response

Collect a file output with FJQSS, fjsnap, or pclsnap for troubleshooting information then contact field engineers.

ERROR: sfcfs_mds:1002: primary MDS is terminated

Explanation

Primary MDS was terminated abnormally.

Response

Collect a file output with FJQSS, fjsnap, or pclsnap for troubleshooting information then contact field engineers.

ERROR: sfcfs_mds:1004: can't execute primary MDS
ERROR: sfcfs_mds:1005: can't execute secondary MDS

Explanation

Startup of the primary MDS or secondary MDS failed.

Response

Check following points.

Quantity of memory is enough.
- Check MemFree and SwapFree of /proc/meminfo.
Connection state of a network is normal.
- Check whether DNS server is operating normally.
- Check whether connection of LAN is normal.
- Check whether network interface card is normal by using ip(8).
Setting of network is normal.
- Check whether all the IP addresses of the nodes that share the file system are correctly registered in host database (such as /etc/hosts). For details, see "11.2.2 Registering information in host database" in this manual.
- Check whether IP address of a DNS server is described in /etc/resolv.conf.
- Check whether /etc/nsswitch.conf is described correctly.
- Check whether fire wall is set up correctly so that access of port used by GFS may not be refused.
State of volume of GDS is normal.
- Check state of volume of GDS by sdxinfo command. For details on sdxinfo, see "Appendix B.1.6 sdxinfo" of "PRIMECLUSTER Global Disk Services Configuration and Administration Guide."
Isn't the disk breakdown?
- Confirm whether the error message of GDS or a disk driver is output to /var/log/messages. When the error message is output, please perform disposal corresponding to the message. For details, see "PRIMECLUSTER Global Disk Services Configuration and Administration Guide" or the manual of disk units.
Is the file system configuration information normal?
- Please execute the following commands and check that the host name of the shared node which the sfcinfo(8) command and the sfcrscinfo(8) command display and the host name specified to be -m option of the sfcadm(8) command execution line obtained by the sfcgetconf(8) command are the same.
  - sfcinfo -n <representative partition of the file system>
  - sfcrscinfo -m <representative partition of the file system>
  - sfcgetconf -d <representative partition of the file system> <output destination file name>
- When not the same, please change the -m option to the -u option of the sfcadm(8) command execution line obtained by the sfcgetconf(8) command, and then execute the command line.

If the problem cannot be solved by the response above, collect the diagnostic data with FJQSS, fjsnap, or pclsnap, and contact field engineers.
For details on commands, see sfcinfo(8), sfcrscinfo(8) and sfcgetconf(8) in this manual. Moreover for the details of the sfcadm(8) command execution line obtained by sfcgetconf(8), see "9.6 Backup of the management partition information" and "9.7 Restoring of the management partition information" in this manual.

ERROR: sfcfs_mds:1006: can't execute MDS

Explanation

Starting up MDS failed.

Response

Collect a file output with FJQSS, fjsnap, or pclsnap for troubleshooting information then contact field engineers.

ERROR: sfcfs_mds:1007: fsid: MDS terminated because logreplay failed. errcode=errcode

Explanation

The log replay by MDS of the file system failed.

Response

Collect a file output with FJQSS, fjsnap, or pclsnap for troubleshooting information then contact field engineers.

ERROR: sfcfs_mds:1008: fsid: Filesystem status is FSBAD, so MDS can not run.fs_clean=fs_clean

Explanation

Starting up MDS failed, because MDS detected an error in file system.

Response

Unmount the relevant file system from all of the nodes, and execute sfcfsck(8) specifying -o nolog.

ERROR: sfcfs_mds:1009: fsid: MDS terminated abnormally because I/O error occurred continuously on meta-data area of the disk.

Explanation

MDS terminated abnormally because I/O error occurred continuously on meta-data area of the disk.

Response

To resume the system operation, execute sfcfsck(8) in the full check mode first.

Then back up the file system and correct the hardware failure by replacing the failed disk with a new one, and restore the backed-up data to recover the file system.