home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
DOS/V Power Report 1997 March
/
VPR9703B.ISO
/
driver
/
novell
/
nwpatch
/
server
/
common
/
libup8j.exe
/
CHANGES.410
< prev
next >
Wrap
Text File
|
1996-04-18
|
2KB
|
83 lines
Changes log for NetWare v4.10 CLIB.NLM version 4.10h.
Symptom:
Server console would hang after performing a backup that required
a server to server connection. After console would hang the server
may eventually Abend or hard hang. This was due to a double
semaphore condition in the server to server connection code.
Solution:
Condition has been fixed.
******
Symptom:
Server would experience a Hang or appear to have High Utilization.
Also when unloading a module the console screen will not release
the console and the module will not unload properly. This was
due to a double semaphore condition.
Solution:
Condition has been fixed.
******
Symptom:
An NLM will fail in the Console Command Process, after putting
SYCHRONIZE in the definition file. and if SynchronizeStart is
never called.
Solution:
We now check that flag.
******
Symptom:
NWDSDeleteUser would not NULL out the TDS pointer after it
freed it.
Solution:
With the new DSAPI it can get freed again. If we NULL it out
it will not doing anything. If you don't it will abend the
machine. Also we needed to check to make sure the NLM was the
owner of the User.
******
Symptom:
Performance problems.
Solution:
Added some semaphoring around the session structures.
Especially in NCPFreeUsersConnections & NCPLogouts.
******
Symptom:
Response times on timeouts is too slow.
Solution:
Added the field "AlreadyTimedOut". When this field gets
set we no longer have to worry about timeout length.
******
Symptom:
Solution:
Modified code in NCPFreeUserConnections() so when we are walk-
ing a connection list with a node that has been deleted we
start over at the beginning. Fixes SPD #102311.
******
Symptom:
Abends could occur if a connection were destroyed when in
certain code paths.
Solution:
NCPReadFile and NCPWriteFile did not semaphore the connection
structure. Added some semaphore checking.
******
Symptom:
There were a couple of paths were a deadlock could occur
with the CacheRoot Semaphore.
Solution:
Corrected these code paths.