Are any additional steps required when configuring multipath storage on a system where /var is a separate file system?

Issue

• LVM commands are hanging on certain devices
• Data corruption is occurring on SAN luns
• The multipath devices are remapping themselves while not being rebooted.  This occurs while system is up and running .

• Have seen that the LUN's presented from the SAN have become read-only with the following error message:

      kernel: EXT3-fs error (device dm-11): ext3_readdir: bad entry in directory #2: rec_len is smaller than minimal - offset=0, inode=0, rec_len=0, name_len=0
      kernel: Aborting journal on device dm-11.
      kernel: ext3_abort called.
      kernel: EXT3-fs error (device dm-11): ext3_journal_start_sb: Detected aborted journal
      kernel: Remounting filesystem read-only
  

note: there is more than 1 cause of the above error

• One problem is with dm-multipath which appears to be mapping to the wrong LUN presented from our SAN.
• Multiple mpaths (mpath10 and mpath11) that both appear to be mapped to lun #10 and this is causing one of the mpaths (mpath 11) shows as empty.
• System was showing 100% iowait, was rebooted and now is having issues activating logical volume.
• Multipath appears to have renumbered the devices and is lost/confused about LVs on those LUNs
• Multipath appears to have mapped the same devices into multiple maps, or is reporting the wrong mpath names for certain devices

• Multipath is seeing duplicate physical volumes after a reboot, corrupting the filesystem.  After a reboot, the following error appears from multipath:

      # multipath -v2 
      remove: mpath53 (dup of mpath55) 
      mpath53: map in use 
      remove: mpath54 (dup of mpath56)
      mpath54: map in use 
  

• Friendly names not working properly, missing /dev/mapper/mpath* names during boot which prevents volume groups being properly recognized and mounted.  The following is seen during boot:

Cannot make directory [/var/lib] : Read-only file system