HP3000-L Archives

May 1996, Week 2

HP3000-L@RAVEN.UTC.EDU

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
Reply To:
Date:
Fri, 10 May 1996 17:39:58 -0700
Content-Type:
text/plain
Parts/Attachments:
text/plain (23 lines)
Randy wrote:
>Be sure that the FLIMIT on the message file is strictly greater than
>the expected number of simultaneous opens against the file.  We have a
>customer whose applications were accessing a message file with an
>FLIMIT of 18.  When the 19th FOPEN occurred, all activity on the
>message file halted.  Not just the 19th process, but _all_ of them.
 
The system reserves one(?) record(block?) in the message file for every
current "writer" of the file.  This ensures that when a writer closes
the file (or aborts) that the system will have a place to write the
"close record" for that writer without the possibility of encountering
a full message file.  If you have a limit of 18, and 18 people with the
file open for write access, then there are no blocks in the file
available for data, and a 19th user will impede waiting for room to
write the "open record" for itself.  As Randy says, never limit the limit
on your message file.  Always leave plenty of room for the (normally
invisible) open and close records.  Especially now that we have Native
Mode Message Files which can have up to 16K(32K?) simultaneous writers,
up from the old limit of 384. (unless they changed the way this works.  I
haven't tested it).
 
G.

ATOM RSS1 RSS2