Hi,
Have you ever try TSA5up9.exe? Maybe it could solve your problem and
you could find it at the Novell site.
PS: Below you could find the fixes of TSA5up9:
Regards
cigdem
NETWARE 5.X
-------------
TSA500.nlm
________
1. Backup of Trustees not happening because of context corruption
In one specific case the TSA created a connection to use with DS in order to
handle file system trustees. This is achieved by the TSA creating a
connection, creating a DS context dependent on that connection then logging
in
and authenticating using that DS context. Trustee backups implicitly use an
associated connection. If, during the backup SMDR disconnects a connection
it
is using that co-incidentally happens to be the same one that TSA associated
with the DS context it is using for trustee backup then that connection is
broken. TSA did not fail a backup if trustees could not be identified for a
file. This has been fixed by separating the TSA from the SMDR context,
making
the TSA report failure to backup trustees and by having the TSA attempt to
re-connect a failed DS context handle. These fixes specifically resolve an
issue where trustee information was not being backed up.
2. NMAS support for TSA
3. ParseDataSet api does not parse the paths correctly.
4. Files when migrated from NW4.x to NW5.x are getting overwritten
5. Backing up a server that does not have LONG namespace and when restored
to a server with DOS and LONG abend is happening.
6. While renaming a volume in a backup which has multiple volumes in a
single session either all is restored or nothing is restored.
7. TaskID to NSS was fixed
8. Restore of read only files was fixed
9. Assumption of contiguous volume ID was fixed.
10. Incorrect passing of lenpreceeded string was fixed.
11. Abend caused because of null string passed to strlen was fixed.
12. 58 chars problem in filename was fixed. 58 hex is a semicolon ascii
char.
The length field had a 58 which confused sms. The file would not be
restored.
13. Fix for defect zCREATE_TRUNCATE_IF_THERE for CSP6.
The defect : An NSS file if it already exists and its size (say20KB) is
greater than the backed up version (say 10KB), then the restore would
only overwrite the first 10KB.
This was fixed by giving the option zCREATE_TRUNCATE_IF_THERE in
createFlags option.
14. Fixes a problem where multiple threads in the same thread group are
logged
out when a volume finishes backing up. Whichever volume finishes backing
up first will cause sms to logout the connection. All other threads in
that
same thread group will be logged out as well. This causes the backup to
fail.
This change was made to the 6/6/01 clib so that multiple connections
would
not be left hanging around after the initial connection for that
thread was logged out.
Symptoms might be the inability to run multiple backup jobs at the same
time.
275 or 259 errors being returned from tsa500.nlm. Or any number of
connection
related errors.
15. Fixed problems restoring mac files to nss volumes. Errors would include
fffdffcc. This means that a file cannot be written to.
16. defect289827 - if the path length is more than 255 char, then an abend
was
observed in smtsarsp_SeparateDataSetName, while browsing thru sbcon.
Path lengths are handled dynamically by adding additional memory.
SMDR.NLM
________
1. IPX dependency of SMDR removed.
2. ParseDataSet api hanging is fixed.
3. Support for specifying an IPX address is supported.
4. Small memory leaks in SMDR fixed
5. fixes a timeout issue due to the timeout value being set too low. This
would
show up in the backup and restore as errors: FFFEFFD4, FFFEFFCC and
FFFEFFB1. The fffeffb1
error is normal to see at the end of a job. It's the smdr.nlm disconnect
request. The problem is if you see it in the middle of a job. These are
connection lost related errors.
6. Some performance enhancements were added.
7. Fixes some abends when establishing a connection to a remote smdr.nlm.
defects: 000267150, 000268157, 000254557
8. Fixed problems of connections getting lost. Then when the new connection
would be established, monitor.nlm would show more connections in a hung
state
that could never be cleared.
9. Fixes a problem where files are opened in deny write mode. An fffdffd6
error
would be returned in NWSMTSReadDataSet. NWSMTSScanDataSetEnd would return
fffefffc errors. This usually happens when backing up a 5.x server to a
4.11
box. If the tape software is configured to open files in the DENY NONE IF
DENY
WRITE FAILS mode, then the problem may occur when the backup hits one of
these files
opened in this manner.
10. Do a remote backup of an NW5.1 sp4 server and simultaneously do a
backup of the host backup server running the tape software.
Stop the job on one server by doing an SMSSTOP. Smdr will abend.
eip is pointing to SMmem_Delete
TSANDS.NLM
__________
1. A memory leak when backing up large nds databases is fixed
2. Tsands would return the error fffdfec7, when backing up nds if the tree
has the same
name as a container.
3. Fixes a problem with error fffdfffc when attaching to the tsands.nlm. The
newer tsands.nlms check for a DNS rooted treename such as t=novell.com
instead
of just t=novell. If it doesn't find one the above error can occur.
FFFDFFFC NWSMTS_CANT_ALLOC_DIR_HANDLE: Cannot allocate a directory handle.
This
does not refer to a file system directory handle. It refers to info from
NDS.
4. Fixes a problem where the backup cannot browse to lower portions of the
nds tree
except the root object.
SBCON
_____
1. Some files can't be restored (DEF255920) This happens mostly when
backingup
from a 4.x server and restoring to a 5.x box. No errors appear, but files
are
missing after the restore.
2. No restored directories are ever written to the Log file.(DEF260175)
3. 254461 - Sbcon backup log shows -2: Server specific info. -2 is shown as
the name space.
4. 254467 - If "Exclude Data Streams" option is selected, log and status
window show the total size backed up as the sum of the sizes of all
files and dirs.
5. 254525 - Abend when a directory is renamed to the max length possible.
6. 254537 - "Backup compressed files as expanded files" appears in the open
mode options screen in Restore.
7. 254627 - Backup status windows and log files are not identical.
8. Provides added security for backup jobs via encryption.
9. Fixes a memory leak caused by doing a selective restore with the option
Include Path Files selected. The memory on the server will slowly drain
away. This option is in the "subsets of what you want to restore" menu.
The
restore is done similar to tid 10025655.
============================================================================
==
SME.NLM
_______
1. Changes have been made to SME.nlm to update the log file only after
the entire file is restored successfully.
2. Fixes a problem where the sme.nlm will always prompt to be unloaded.
the job cannot be run unattended.
If unloaded, it abends the server, leaving it loaded causes the job to
fail,
with a status of "could not start".
============================================================================
==
NETWARE 4.X
___________
DSBACKER.NLM - Fixes a problem where an "insufficient buffers" error would
arise when backing up the server specific information.
============================================================================
==
TSACORE.DLL
___________
1. fixes the issue of the modified time getting changed when data is backed
up
from NT and restored to NetWare.
2. Fixes a problem backing up files from NT and restoring them to a netware
5.1
server. Any file with a character that has a 5c hex value (this is a
backslash)
would not restore to the netware box. There was also a similar problem fixed
in nlmlib.nlm. This is part of clib. This was mostly seen on japanese
servers.
tsamain.exe and tsaresp.dll are also included in this fix.
============================================================================
=
-----Original Message-----
From: James Edwards [mailto:jedwards AT SOS.STATE.TX DOT US]
Sent: Thursday, August 01, 2002 6:07 PM
To: NETWORKER AT LISTMAIL.TEMPLE DOT EDU
Subject: Re: [Networker] Novell abends
Simon, I don't know what help this will be but we used to have a constant
problem with our Novell servers abending because of Legato. I could almost
always make it happen if I made it search for a server from the GUI on the
client. It has been so long ago that I can't recall exactly what fixed it
but I seem to remember that it was a combination of the latest Legato client
and several of the Novell networking NLMs. Legato can be very picky about
which versions of what NLMs work but tech support should be able to give you
the list of required NLMs and what version works best.
Sorry to be so vague but it was long ago and I'm old.
Hope this helps.
Jim Edwards
Systems Manager
Texas Secretary of State
-----Original Message-----
From: Simon Phillips [mailto:sdph AT DML.AC DOT UK]
Sent: Thursday, August 01, 2002 9:04 AM
To: NETWORKER AT LISTMAIL.TEMPLE DOT EDU
Subject: [Networker] Novell abends
Hi,
I'm running Networker 4.2.1 client on Novell, backing up to a Sun with
Networker 6.1.1.
Several servers abend, all with the following messages
Console Log:
SNAPSHOT.NLM:*Trying to mount volume !MAIL at 4:50:26 7/28/2002. SRV1 <1>:
SRV1 <1>:mount !MAIL Mounting Volume !MAIL
** Reading In FAT
** Verifying Directory FAT Chain
** Scanning the Directory
** Checking Subdirectories
** Scanning Files with Trustee Nodes
** Scanning Deleted Files
** Linking the Name Spaces
** Shrinking the Extended Directory
Problem with file 0728POA.001, length extended.
Problem with file 0728POA.001, old length = 0, new length = 65536 Problem
with file 0728POA.001, length extended. Problem with file 0728POA.001, old
length = 0, new length = 65536 Problem with file POOL0003.DAT, length
extended. Problem with file POOL0003.DAT, old length = 0, new length =
13107200 Problem with file POOL0002.DAT, length extended. Problem with file
POOL0002.DAT, old length = 0, new length = 131072 Problem with file
POOL0001.DAT, length extended. Problem with file POOL0001.DAT, old length =
0, new length = 2097152 Problem with file POOL0000.DAT, length extended.
Problem with file POOL0000.DAT, old length = 0, new length = 131072
SNAPSHOT.NLM: Volume !MAIL mounted at 4:51:29 7/28/2002.
SRV1 <1>:
SRV1 <1>:dismount !MAIL
SNAPSHOT.NLM: Volume !MAIL dismounted at 5:49:49 7/28/2002. Dismounting
volume !MAIL Volume !MAIL has been dismounted.
SNAPSHOT.NLM: Deleting snapshot virtual device for !MAIL...
28-07-2002 5:50:19 am: NWPA-3.6-0
NWPA-004: The CDM driver deactivated device [V5E0-A2-D0:0] Snapshot of
#
0 IBM RAID rev:1 due to a device failure.
28-07-2002 5:50:19 am: SERVER-5.0-1534 [nmID=B0013]
Device "[V5E0-A2-D0:0] Snapshot of # 0 IBM RAID rev:1" deactivated by
driver due to device failure.
SNAPSHOT.NLM: FINISHED deleting snapshot virtual device.
_ _ _ _ _ _ _
SNAPSHOT.NLM:*Trying to mount volume !SHARED at 9:13:23 8/01/2002.
SRV1:
SRV1:mount !SHARED
Mounting Volume !SHARED
** Reading In FAT
** Verifying Directory FAT Chain
** Validating the Extended Directory
** Scanning the Directory
** Checking Subdirectories
** Scanning Files with Trustee Nodes
** Scanning Deleted Files
** Linking the Name Spaces
** Shrinking the Extended Directory
Number of unclaimed FAT blocks that were freed 148
Problem with file ALLSTU.LDB, length extended.
Problem with file ALLSTU.LDB, old length = 0, new length = 65536 Problem
with file ~$02_F~1.DOC, length extended. Problem with file ~$02_F~1.DOC,
old length = 0, new length = 65536 Problem with file INDUCT~1.LDB, length
extended. Problem with file INDUCT~1.LDB, old length = 0, new length =
65536
SNAPSHOT.NLM: Volume !SHARED mounted at 9:14:28 8/01/2002.
The running process will be suspended.
1-08-2002 9:14:29 am: SERVER-5.0-4631 [nmID=1001C]
WARNING! Server SRV1 experienced a critical error. The offending
process was suspended or recovered. However, services hosted by this
server may have been affected.
Abend Log
Additional Information:
The CPU encountered a problem executing code in SERVER.NLM. The problem
may be in that module or in data passed to that module by a process owned by
SNAPSHOT.NLM.
____________________________
Has anybody seen this kind of fault before? If so, is there a solution??
Will summarise.
Simon Philllips
System Administrator/DBA
UHI Millennium Institute
Dunstaffnage Marine Laboratory
Dunbeg, Oban, Argyll
PA37 1QA
Tel: +44 (0) 1631 559286
Mobile: 07885-663308
Fax: 01631 559001
Email: s.phillips AT dml.ac DOT uk
http://www.uhi.ac.uk
--
Note: To sign off this list, send a "signoff" command via email to
listserv AT listmail.temple DOT edu or visit the list's Web site at
http://listmail.temple.edu/archives/networker.html where you can also view
and post messages to the list.
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
--
Note: To sign off this list, send a "signoff" command via email
to listserv AT listmail.temple DOT edu or visit the list's Web site at
http://listmail.temple.edu/archives/networker.html where you can
also view and post messages to the list.
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
--
Note: To sign off this list, send a "signoff" command via email
to listserv AT listmail.temple DOT edu or visit the list's Web site at
http://listmail.temple.edu/archives/networker.html where you can
also view and post messages to the list.
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
|
