Kbase P116612: Remote server crashes with client error 438 and error 49
| Autor |
Progress Software Corporation - Progress |
| Acesso |
Público |
| Publicação |
8/12/2010 |
|
Status: Verified
SYMPTOM(s):
Remote server crashes with client error 438 and error 49
After an error 438 there is a memory violation 49 bringing down remote server(s)
Client attempting to access recid <record-number> not part of file <file-number>. (438)
SYSTEM ERROR: Memory violation. (49)
Indexes for table listed in error 438 are clean
ASCII dump of table listed in error 438 successful
FACT(s) (Environment):
Progress 9.1x
UNIX
Compaq Tru64 UNIX
PA-RISC 64-bit
ia64 (Itanium)
PowerPC 64-bit
x86_64 (AMD64/EMT64)
ia64 (Intel Itanium 64-bit)
IBM Power (64-bit)
Sun Solaris SPARC 64-bit
Sun Solaris Intel 64-bit
CAUSE:
Bug# OE00123769
CAUSE:
The error 438 is due to a failed find. It could also be that RECIDs are stored in the database or called in the application code. If such a RECID was used for a FIND, and it was no longer valid due to the database being changed or records simultaneously being deleted for example, then the 438 could result.
The fact that this instance is followed by error 49 and remoteserver stopping is where the Progress bug issue lies. The impact of bringing down a remote server, could be potentially dangerous when the "other" C/S connections to that remote server are also interrupted, perhaps resulting in an abnormal shutdown.
Todate, this issue has not been reproducible outside of production environments and in these environments reported cases where this also results in the broker terminating, have been rare. The common denominator appears to be the 64-bit Solaris and 64-bit Progress for Solaris platforms.
FIX:
This issue has been addressed in OpenEdge 10.0B05 and 10.1B