How do I regulate short-lived documents development in SQL Anywhere?

We are presently experiencing an expanding short-lived documents trouble. Seeing among our websites we can see a development of 100 - 200 MEGABYTES a day at one website we are seeing. At this website a failing took place when the temperature documents got to 20Gb and also a vacuum concern took place.

We are presently subduing timeouts. - ti and also - tl are readied to absolutely no. What is the opportunity that temperature tables are accumulating as a result of this arrangement?

In addition. To better my understanding of the - tl flag, is the adhering to a real declaration: The link will certainly not be reset unless the customer can not be gotten to using tcpip.

0
2019-05-18 20:58:05
Source Share
Answers: 2

You are appropriate that the - ti and also - tl buttons are unconnected to temp room.

Pertaining to the liveness concern, both the customer and also web server send liveness packages every n/3 secs, where n is the liveness timeout value. This only takes place if nothing else packages have actually been sent out in the meanwhile. If either side has not receive any kind of packages from the opposite side after n secs, the link is gone down.

Liveness is not essential in case where the link is in fact come by the opposite side (eg. the application/server collisions or the equipment restarts) given that the TCP link will certainly be shut quickly. Liveness serves, nonetheless, for identifying applications that are installed and also network troubles that protect against packages from surviving yet do not create the TCP link to be gone down.

0
2019-05-21 10:56:50
Source

It is not likely that either - ti 0 or - tl 0 have anything to do with the temperature room trouble in SQL Anywhere. It is more than likely the outcome of a runaway question. If you are making use of variation 9 you can activate restriction monitoring as adheres to:

SET OPTION PUBLIC.temp _ space_limit_check = 'ON' ;

In variations 10 and also 11 that alternative need to currently get on, yet possibly it obtained switched off. The new max_temp_space alternative is additionally valuable.

In earlier variations, you'll simply need to locate the runaway questions ; Foxhound might aid: http://www.risingroad.com/foxhound/index.html

See additionally "Danger! The Queries are Stampeding!" at http://sqlanywhere.blogspot.com/2009/03/danger-queries-are-stampeding.html

FYI the - ti 0 endless lack of exercise timeout setup is really usual when you anticipate extended periods of lack of exercise ; as an example, over night on a large internet - based link swimming pool.

Nonetheless, the - tl 0 endless liveness timeout setup is in fact unsafe if it causes great deals of zombie links accumulating (the customers are lengthy dead yet the web server holds the links open). Like the Help claims, it is generally far better to increase the timeout duration if you are having early liveness timeout troubles ; as an example, - tl 3600 for one hr.

AFAIK the declaration "The link will certainly not be reset unless the customer can not be gotten to using tcpip" seems an over - simplification: the monitoring of liveness packages appears to be instead extra intricate than a straightforward "can not be gotten to" examination.

Breck

0
2019-05-21 02:36:09
Source