[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference iosg::all-in-1_v30

Title:	OLD ALL-IN-1 (tm) Support Conference
Notice:	Closed - See Note 4331.l to move to IOSG::ALL-IN-1
Moderator:	IOSG::PYE

Created:	Thu Jan 30 1992
Last Modified:	Tue Jan 23 1996
Last Successful Update:	Fri Jun 06 1997
Number of topics:	4343
Total number of notes:	18308

745.0. "DATA_FILE lock timeout query" by COMICS::TALBOT (Trevor Talbot) Fri May 22 1992 12:31

    Hi,
    

	I wonder if you could help me with an answer to this question? If I 
    have an entry form that has a .FILE using an FDL file to describe the 
    keys etc. and within the FDL I have this

	CONNECT
		timeout_enable   YES
		timeout_period   8

	If I call a script which makes use of the DATA_FILE lock dset/on and 
    then from another session(but at the same time) I call it again, one gains 
    access to update the record, the other doesn't. The same is true if the 
    entry form is used from both sessions and then update on the same record. 
    This is the expected behaviour. The timeout period, set in the FDL for 8 
    secs is not obeyed, but rather a default value of 30 secs seems to be in 
    effect. I need to know if the timeout period can be changed, and how, or 
    if not, why? Is it an RMS restriction or a designed use of RMS by ALL-IN-1?
    
    Thanks for your help

    -Trev

T.R	Title	User	Personal Name	Date	Lines
745.1	ALL-IN-1 takes 6 - 5 second waits	SHALOT::WARFORD	Richard Warford @OPA DTN 393-7495	`Fri May 22 1992 14:01`	5
	ALL-IN-1 has a hard coded 30 second delay loop trying to get access to a record in an RMS file. Sorry - without a patch I don't believe it is possible to change. Rick
745.2	I just couldn't resist!	SHALOT::DUNCAN	Joe - CIS/EIC Doc. Mgmt. Solution Set Consultant	`Fri May 22 1992 15:50`	5
	Of course, if the data file were actually an Rdb table (or even an RMS file access via RdbAccess for RMS), and you were using SRA to access the data, you have control over the timeout interval used! ;-) Joe Duncan @ OPA
745.3	there is no 30 second timer	SKNNER::SKINNER	I'm doing my EARS	`Fri May 22 1992 16:42`	13
	Actually, the I/O timer for a locked record is 1 second, and is retried up to 10 times for a maximum of 10 seconds of delay from ALL-IN-1's I/O routines. If RMS "imposes" a delay timer, then that delay (times the 10 retries) would be added to ALL-IN-1's. In looking at the VMS File Definition Language Facility manual (using Bookreader) under CONNECT, it appears that you must set WAIT_FOR_RECORD before the TIMEOUT_ENABLE/_PERIOD is honored. Did you? BTW, the Bookreader version of this manual has the TIMEOUT_ENABLE/_PERIOD descriptions written very poorly, as it implies that _ENABLE is a numeric attribute (it's not). I have no idea if the paper version of the manual is different than the Bookreader version. /Marty
745.4	I'm sorry; there is	IOSG::SHOVE	Dave Shove -- REO-D/3C	`Fri May 22 1992 16:57`	6
	But there's also a bit of code in OARMS which explicitly sets the value for the RMS timeout to 30 secs. As in: RAB[RAB$B_TMO] = 30; Dave.
745.8	sorry for any misled readers	SKNNER::SKINNER	I'm doing my EARS	`Fri May 22 1992 17:33`	22
	I stand corrected. I didn't search the sources for the TMO value, but tried to relay information I knew. My information was stale, and only related to calls to the OAIO module, not OARMS. There are no retries being done by ALL-IN-1 in the OARMS module, so whatever value is stored in RAB$B_TMO is the only delay. But looking at what has now been pointed out to me, how does the TMO value of 30 interrelate with .0's question? I presume that his explicit setting of 8 seconds is not considered. If so, why was this approach taken? If the file "designer" can't get the desired features of the .FDL to be used, how can future versions of ALL-IN-1 be coded so that they can? Perhaps by conditionalizing the setting of RAB$B_TMO to only happen if the file doesn't have an explicit setting? One thing seems to be "missing" from ALL-IN-1: a central module/linker options for setting "limiting" values such as this choice of 30, the number of SDAFs and... Not that customers should feel encouraged to change these things, but at least they could be patched more easily if the need arises as they would be forced to have global names for the locations. /Marty
745.10		COMICS::TALBOT	Trevor Talbot	`Wed May 27 1992 18:03`	5
	Officially logged, SPR number is R11-51871 ...just in case you need to follow up. -Trev