bishop@ecsvax.UUCP.UUCP (05/23/86)
A few weeks ago I posted a description of problem that we've been having
dealing with processes hanging in RWAST. If you didn't get the article (I
didn't at my site), the original article is at the end of this.
For those of you having the same problem, it looks like you get to live
with it. I got no useful replies. Just lots and lots of "I've been having
this problem too. Let me know if you find anything." One person thought
that there was a SYSGEN parameter dealing with a timeout value, but didn't
say what it is (and hasn't answered any of my mail) (and I didn't see it
described in SYSGEN HELP). If anyone ever finds a solution, it would
be nice to see it posted. There are apparently several causes, coming from
various types of waits. One person's NETACP hopped into RWAST and stayed.
Did wonders for his system. Oh well.
---
Description of problem follows:
We are having trouble with terminals hanging. The process
associated with the terminal goes into RWAST and stays. The process
hangs around, the user does not show up on a SHOW USERS, and anyone who
does not have privilege on the process gets the expected privilege
violation error when trying to access it, but those who should have
access get a non-existent process error.
The program that does this uses global memory, common event flag
clusters (both group wide, not system), and timer queue ASTs. It
appears to be linked to instances where the process that created the
global section exits, but others still have channels assigned to the
global section. It does not happen all the time.
We are running VMS 4.2, no DECNET, and non-privileged accounts. Has
anyone seen/had this problem? Any suggestions or fixes?
Please respond via mail. Will be glad to send copies of responses
to any interested people.
--
Alan Bishop
...!decvax!mcnc!ecsvax!bishop
bishop@ecsvax.{UUCP,BITNET}