JRESCUER FAQ

December 14, 2005

This README contains frequently-asked questions 
related to JRescuer (TM) for OS/2 and eComStation


CONTENTS
________

   Q1: JRescuer crashes on recovering of the N-th file
   Q2: What are the requirements of JRescuer if no JFS.IFS loaded
   Q3: How to make JRescuer process linux JFS volumes?
   Q4: I am lazy to read documentation. How to recover files from x: volume?
   Q5: Can I use JFS filesystem on USBMSD devices? 
   Q6: The root directory is not available. How to recover files?
   Q7: Can chkdsk intensify the state of damaged JFS volume?
   Q8: How to recover a file from JFS which is now 0 bytes long? 
   Q9: Which value of /CACHE:x  to use for JFS.IFS?
  Q10: Chkdsk returns error JFS0138
  Q11: Why JRescuer doesn't restore original names of files?
  Q12: No valid superblocks.
  Q13: Trap 3.
  Q14: Is there an ifs for Windows to read jfs?
  Q15: Can you save the fullpath of file to .LONGNAME? 
  Q17: Can I use Defragfs.exe for JFS?    
  Q18: JRescuer returns error "INODE: Error reading iag"
  Q19: I have removed a file one minute ago. JUne didn't found it!  
  Q20: Why JRescuer and JUne don't lock the volume while processing it?
  Q21: JRescuer returns error "##### NotConventions", where ##### is some number.
  Q22: JRescuer returns error "##### NotConventions" for all files and directories.  
  Q23: Error: GetXtree: Error create file Restored.From.JFS, rc=5 action=65535
  Q24: JUne traps on spanned JFS disk.
* Q25: Why the size of swap file on JFS can't exceed 2 Gb?
* Q26: I copied small icon from network drive to E:, removed it. JUne can't find the file. 
* Q27: Why do you work on JRescuer? Concentrate on JFS.IFS improvement!


   
Q1: JRescuer crashes on the recovering of N-th file
A1: Use lst2cmd.cmd script to generate command file which will recover
    files one by one.
    To generate list use command: Jrescuer x: /d >list.lst   
    
Q2: What are the requirements of JRescuer if no JFS.IFS loaded
A2: JRescuer works fine without additional dll or ifs.
    

Q3: How to make JRescuer process linux JFS volumes?
A3: JRescuer allows process volumes without assigned letter. Use -d switch.


Q4: I am lazy to read documentation. How to recover files from x: volume?
A4: Usually this command is enough:
      Jrescuer x: /O=y:\
    where y: is a volume with free Hard Drive space for recovered files
   

Q5: Can I use JFS filesystem on USBMSD devices?
A5: Yes you can format flash cards as JFS if using JFS for eComStation 
    (JFS package released in 2005 or later via eCS beta-zone). 
    Attention: if you are using lazywrite (/LW:ON) then eject x: command 
    may require some minutes. Don't disconnect device while eject! 
    Example: if using an old flash-card actively and then write big 
    file (128 Mb), the eject will require more than 10 minutes.

   
Q6: The root directory is not available. How to recover files?
A6: Follow this steps:

     1) Start JRescuer with /D switch: JRescuer x: /D > y:\Ddump.log

     2) Start JRescuer with /S=2 switch: JRescuer x: /S=2 > y:\S2dump.log
          If JRescuer finds only inodes mentioned in Ddump.log then 
          start JRescuer /S=N for every inode N mentioned in Ddump.
      
     3) Follow the overview, readme, etc.   
   
    
Q7: Can chkdsk intensify the state of damaged JFS volume?
A7: Chkdsk may correct the volume so JRescuer fails on recover.
    Nevertheless the data may be saved with additional efforts.

     
Q8: How to recover a file from JFS which is now 0 bytes long because  
    of dirty shutdown
A8: Try JUne and search necessary file among removed. No guarantee. 
    The probability of success is 80%
     
    
Q9: Which value of /CACHE:x  to use for JFS.IFS?
A9: The size of cache is measured in KBytes. You can use 4 Mb .. 200 Mb 
    (4096 .. 204800)
    There`s no point in setting less than 20 Mb for JFS cache.
    
    
Q10: Chkdsk returns error JFS0138
A10: It's normal. Every mounted drive is "ubnormal" for chkdsk.

    
Q11: Why JRescuer doesn't restore original names of files?
A11: JFS filesystem doesn't keep names of removed files.
     Please execute the script setln.cmd to save names of existing 
     files to .LONGNAME EA 
     Run it periodically for all files in specifiled directory 
     and subdirectories.
     Default recommendation: Run it 2 times per week for all JFS volumes.
    

Q12: No valid superblocks.
     Is there any way to rescue files from a JFS partition that has no
     valid superblocks?  Is there a way to "restore" or repair invalid
     superblocks?  If a partition was corrupted, is there a way to search
     every sector of a hard disk to locate a misplaced superblock?
A12: Check yourself:
      * Are you sure that both superblocks was damaged? ISJ allows select 
        superblock. ISJ is distributed with JRescuer.

      * If no superblock, then JRescuer uses default parameters. default size 
        of cluster = 4096. Did you changed this parameter before format?

      * try search all inodes
  
  
Q13: Trap 3. Validfs.exe runned after chkdsk /f  doesn't report that 
     "fs is valid". It reports which inodes have violation in structure. 
A13: Determine the name of file using JRescuer. Save it, remove old one.
     Continue until Validfs informs that "fs is valid"

     
Q14: Is there an ifs for Windows to read jfs?
A14: No

      
Q15: Can you save the fullpath of file to .LONGNAME? 
       (to restore file and its directory in the future)
A15: The WPS uses .LONGNAME and expects to see name only, not fullpath.


Q16: How many disk space the .LONGNAME requires?
A16: An inode occupies one cluster per file/directory. The default size 
     of cluster is 4096 bytes. The size of inode itself is 512 bytes. 
     So, there is space 4096-512 bytes for EA and ACC. The max length 
     of LONGNAME is 255 bytes. 
   
     As you can see, the .LONGNAME attribute requies additional:
          0 bytes -- if file doesn't have EA attributes,
          0 or 1 cluster -- if file has EA but doesn't have .LONGNAME 
          
          
Q17: Can I use Defragfs.exe for JFS? 
A17: Don't use it. It is dangerous.


Q18: JRescuer returns error "INODE: Error reading iag"
A18: No inode with such number.


Q19: I have removed a file one minute ago. JUne didn't found it!
A19: Try search the file among existing files. 
     Run
       Jrescuer.exe e: /g=\Treasure\DirectoryOfRemovedFile
     (without drive letter)
     Usually this is enough.

     For example, JRescuer reports:
       Finding dir \Treasure\DirectoryOfRemovedFile\ inode: 303471

    then run 
       jrescuer /u=303471
    to view contents pf the directory
       and /i=303471 
    to recover files
    
    // this answer is not confirmed by the developer
    
    
Q20: Why JRescuer and JUne don't lock the volume while processing it?
A20: There is no need to do this because this applications don't write
     to the volume. Close all applications which may write to the JFS volume. 
  
     
Q21: JRescuer returns error "##### NotConventions", where ##### is some number.
A21: This occurs when unicode is unable to convert name of file to system 
     codepage. Try recover file without name using key 
        Jrescuer x: /U=inode 
     (or /i=inode if this inode is DIR)   
   
     ##### NotConventions occurs: 
      - if your unicode system doesn't work right (the message appears 
        for every file with NLS character in filename)
      - if filename is damaged on physical disk

     
Q22: JRescuer returns error "##### NotConventions" for all files and 
     directories, where ##### is some number.
A22: Check the presence of unicode.sys driver in CONFIG.SYS and 
     presence of system \language directory


Q23: Error: GetXtree: Error create file Restored.From.JFS, rc=5 action=65535
     The program has been working great until now.  I keep getting the 
     following error: 
        GetXtree: Error create file Restored.From.JFS, rc=5 action=65535
A23: This means that Restored.From.JFS exists in the directory. 
     Recover files to other directory.

   
Q24: JUne traps on spanned JFS disk.
A24: Current version of JUne (JRescuer) doesn't support spanned disks.
     New version is under construction.
     
     
Q25: Why the size of swap file on JFS can't exceed 2 Gb?
A25: Research the VIRTUALADDRESSLIMIT= line in CONFIG.SYS
     If you need bigger swap file then set value greater than 2048
     
     
Q26: I copied small icon from network drive to E:, removed it. JUne can't find
     the file. 
A26: It was copied to cache. Then you removed it from the cache.   
     So.. it is missing on the harddisk.
   
     
     
Q27: Why do you work on JRescuer? Concentrate on JFS.IFS improvement!
A27: JRescuer is needed because:
      * there are troubles with hardware
      * traps & other events damage data on disk.
   
     It's necessary work on both directions: recover tools and filesystem driver. 
   
     