home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
ftp.seagate.com
/
2014.07.ftp.seagate.com.tar
/
ftp.seagate.com
/
pub
/
palindrome
/
technote
/
tn9305.asc
< prev
next >
Wrap
Text File
|
1995-03-21
|
11KB
|
237 lines
Tech Note Number: 9305
Date: August 02, 1993
Subject: Troubleshooting 811 & 2000 Series
Errors
Keywords: 811, 2004, 2029, 2030, 2031, 2032,
2035, 2036, 2037, 2123, 2199, 2215,
2231, 2237, Database Integrity
Errors
Network Archivist Version: 2.0 and above
Palindrome Backup Director: 2.1
Troubleshooting 811 Errors
Error 811 "Database Integrity Error" occurs when there is a problem
with the internal structure of the database or there is an
environmental problem that prevents the database from being used
properly. The following Tech Note summarizes procedures for resolving
these errors.
An 811 error is usually preceded or followed by a 2000-series error
which represents the error code generated at the database level. The
following table lists each of the most common database errors and the
appropriate action to take. Following the table on the next page,
suggested actions are listed numerically from #1-#5. If the error you
are receiving is not in the following table, perform all of the
suggested actions from #1-#5 until the problem is resolved.
ERROR SUGGESTED ACTIONS
811 by itself #1 or see "Older NetWare shell versions" below
2004 #2,4,5; use TESTNDX*
2029 #1 should fix -- tag ALL volumes
also see Tech Note TN93-003.ASC "Troubleshooting
2029 Errors"
2030 #1 should fix
2031 Use SNAPDEL*
also see Tech Note TN93-002.ASC "Troubleshooting
2031 Errors"
2032 #2,4,5; use TESTNDX*; try SNAPDEL /K (killswitch)*
also see Tech Note TN93-001.ASC "2032
Errors"
2035 (seek error) #1-#5; also see "Network Problems" below
2036 (read error) #1-#5; use SNAPDEL*; also see "Network Problems"
below
2037 (write error) #3 only (especially #3a); also see "Network
Problems" below
2123 #1-#5
2199 #1 should fix
2206 See 2037 error
2215 #2-#5; use TESTNDX*; also see "Network Problems"
below
2231 #2-#5; use TESTNDX*; also see "Network Problems"
below
2237 #2-#5; use TESTNDX*; also see "Network Problems"
below
*See Database Diagnostic & Repair Tools below
To determine which volume is receiving the database error, there are
several methods. The easiest way is to use the undocumented timestamp
switch (/TS) when running backups:
TNA2TAPE /A /Q /TS
This will create a file called TMPSTAMP.LOG in the installation
directory that shows each operation performed on each volume in
sequence. If an error aborts the backup, examine TMPSTAMP.LOG to see
what volume the last operation was performed on.
Another way to determine which volume is receiving errors is to examine
the display (not tape) journals located in the Display menu. Locate the
last volume listed in the display journals (the journal entry with the
most recent time stamp). This indicates the last volume where a backup
completed. The database errors most likely occured on the volume that
normally follows this volume during a regular backup. Examine the
protected volume list to see which volume follows the last volume
listed in the journals. It is also possible but less likely that the
errors occurred on the last volume listed in the journals.
ACTIONS:
Perform the following actions in the order noted for each database
error listed above. If the database errors you are experiencing are not
listed in the table above, perform all the actions listed below in the
order presented until the problem is resolved. To insure backups of
other volumes are completed, move the volume suspected of corruption to
the bottom of the protected volume list.
1) Verify file history database -
- Run TNARECOV. For users of 2.0f, make sure you are using the
version of TNARECOV dated 10-23-92. Download from the BBS if
necessary.
- Tag volume(s) suspected of being corrupt. If you are unsure
which volume to tag, tag them all.
- Select "Verify File History Database". If verification loops or
hangs, see below.**
- Run backup or use TESTNDX (if appropriate) to determine success.
2) Restore file history database from tape -
- Journal tapes (TNAUTIL /J) and locate an older checkpoint
fileset for the volume. Record the fileset descriptor (e.g.
CP1234).
- Restore file history for server volume S/V: with the following
command line:
TNA2DISK /RH /DCP1234 S/V:
- Run backup or use TESTNDX (if appropriate) to determine success.
- If unsuccessful, repeat this procedure with 1 or 2 older tapes
or until TESTNDX shows zero errors.
3) Eliminate environmental problems -
a) File system problems:
- Free sufficient disk space (at least 2% volume space
remaining).
- Work around invalid ownership (N/A) or disk problems by
making copies of the AV*.PAC files and renaming the
originals (e.g. AV*.OLD). Use the copies.
b) Local (DOS workstation) problems:
- Clean AUTOEXEC and CONFIG (load no unneeded
TSRs/drivers/mem mgrs).
- FILES = 20
- Free up at least 520K of RAM (after loading TNASTART).
- Move backup to alternate PC with different NIC on different
cable.
c) Network problems:
- Look for possible causes of dropped connections.
4) Restore older control database and all histories -
- Rename AS*.PAC to AS*.OLD.
- Insert an older tape (from before problems began).
- Run TNARECOV. This will automatically restore the control
database from the mounted tape.
- Tag all volumes in the protect volume list and select "Recover
File History Database".
- Run backup to determine success.
5) Create new history (and rebuild with Learn-a-tape) -
- This operation is the last resort. Data on tapes must be restored
using command lines unless history is rebuilt using Learn-A-TAPE.
Contact Palindrome Tech Support for more information or
assistance.
- Rename AV*.PAC to AV*.OLD on corrupt volume.
- Delete corrupt volume from protected volume list.
- Reinsert same volume back into protect volume list and select
"Create New History".
- If desired, dump tape library (in utility menu) and use the
Learn-a-tape* utility to rebuild history.
**If TNARECOV/VERIFY loops or hangs perform the following procedure (to
recognize a loop, the version record counter will continue counting
past the target value without stopping) -
- Run TESTNDX.
- If TESTNDX shows there are errors, then proceed to Action #2
above and repeat until TESTNDX shows zero errors.
- If TESTNDX shows NO errors at all, run the DUMPHIST* utility to
repair the database.
DATABASE DIAGNOSTIC & REPAIR TOOLS:
TNARECOV:
The verification routine in versions 2.0f and 2.0g of TNARECOV can fix
a large number of database corruptions. When TNARECOV fixes something,
it will often (but not always) post messages to the log.
Versions of 2.0f TNARECOV dated earlier than 10-23-92 should be
upgraded to this version. Download RCV20FB.ZIP from the Palindrome BBS
(708-505-3336, 8-N-1). Versions of TNA prior to 2.0f should download
the FS_CHK utility. Contact Palindrome Tech Support for details
about this utility.
TESTNDX:
The TESTNDX utility can be used to spot inconsistencies in the database
index file. It is only a diagnostic tool, not a repair tool, so it can
only be used to identify a corrupt database. It is most useful for the
errors listed above that suggest using TESTNDX as part of their
suggested action.
Download TESTNDX.ZIP (PW=Ctree) from the Palindrome BBS. Run this
utility on a corrupt database to see if it has index errors. If it
does, then use TESTNDX to test restored databases until a clean set is
found (see Action #2).
SNAPDEL:
The SNAPDEL utility is used to repair only two errors: 2031 errors and
some 2036 errors. The 2036 errors that are repairable can be identified
by a stack Length (see error log debug information) of "0x35a". Download
SNAPDEL.ZIP from the Palindrome BBS and also download Tech Note TN93-002.ASC
"Troubleshooting 2031 Errors" for more details.
DUMPHIST:
The DUMPHIST utility is used to repair corruptions that can lead to
hangs or loops in the database. It can fix most but not all database
loops, and TESTNDX can be used to spot the others. Use them together to
resolve a looping problem in the database. Looping database
corruptions usually manifest as counters that increment without
stopping in TNA2TAPE or TNARECOV. Run the utility with the following
command line and examine the DUMPHIST.OUT file to see if fixes were
made:
DUMPHIST /CCS /WF /B
Download 2413.ZIP (PW=Peas) for users of TNA 2.0d or later. For
earlier versions of TNA, contact Palindrome Tech Support for earlier
versions of DUMPHIST.
LEARN-A-TAPE:
To rebuild history based on the data in the tape library, use the
Learn-a-tape utility. This utility reads the output files generated by
Dumping the tape (TNAUTIL /DT) and rebuilds the database from the data
in the dump files. Download the file 6057.ZIP to rebuild all or part of
a 2.0f or 2.0g database. Documentation is included. Contact Tech
Support for earlier versions.
ENVIRONMENATAL CAUSES:
Network problems:
Certain repetitive database errors have indicated nework problems in
the past: 2035, 2036, 2037, 2215, 2231, 2237. If these errors occur
repeatedly even after creating new databases, there is probably some
nework problem that needs to be resolved before normal database
operations can continue.
Older NetWare shell versions:
Versions 3.22 and 3.26 of NETX can, under certain circumstances, cause
an 811 error. This error is NOT preceded by any 2000-series error. The
above actions will have no effect on this error. Upgrading the shell
version to 3.31 or higher will resolve the problem. The older shells
were also known to cause other strange problems such as persistent 807
errors, so it is recommended that you upgrade to the most recent shell
version regardless. The most recent shells can be downloaded from
Compuserve or the Palindrome BBS (708-505-3336, 8-N-1).
If you have any questions about these procedures, please contact
Palindrome Tech Support at 708-505-3300. Database corruptions are not
normal. Any installation experiencing recurring database problems
should investigate the cause of these problems.