Are there recommendations for connection retry -ct , are there reasons not to use -ct 1 ? - Forum - OpenEdge Deployment - Progress Community

Are there recommendations for connection retry -ct , are there reasons not to use -ct 1 ?

 Forum

Are there recommendations for connection retry -ct , are there reasons not to use -ct 1 ?

This question is answered

The documentation states -ct has a default of 50 and a minimum of 1.

In the pkb I found that in v8 the value used to be 5 or 10.

I'd like to set the parameter as low as possible but without introducing new issues.

Are there drawbacks to setting this value to 1 ?

What are the situations in LAN in which the first connection would fail and a subsequent try to succeed ? other than somebody starting the database server in the 50 seconds it takes to retry 50 times.

https://documentation.progress.com/output/ua/OpenEdge_latest/index.html#page/dpspr/connection-retry-attempts-(-ct).html

Verified Answer
  • networks are much more reliable now than when this was implemented in version 4. coonnections on a LAN hardly ever fail due to sporadic network dropouts or message corruption. so set it to 1 or 2. if two retries dont work, more won't either.

    (in 30 years, i've never seen anyone actually use this parameter)

All Replies
  • networks are much more reliable now than when this was implemented in version 4. coonnections on a LAN hardly ever fail due to sporadic network dropouts or message corruption. so set it to 1 or 2. if two retries dont work, more won't either.

    (in 30 years, i've never seen anyone actually use this parameter)

  • I didn't know the parameter existed, I was under the impression we needed to change some OS TCP/IP parameter to make connect fail faster.

    I  have some use cases for it.

    • Development environment that is not always up & running.
    • Secundairy databases that are not vital
    • even for the vital databases it seems like a good thing if connection failure is reported faster
  • > On Jun 13, 2019, at 12:15 PM, Gus Bjorklund wrote:

    >

    > so set it to 1 or 2.

    someone in development will have to verify that this parameter actually does something. when orgiginally implemented there were a bunch of other network protocols and several flavors of tcp/ip. now only the standard tcp/ip remains.

  • I can confirm that -ct still does something. As mentioned above, if not set, the default is 50 tries. I'm guessing that -ct stands for "connection try", but functionally, it really is "connection retry". The AVM attempts the connection and then retries N times, where N is set via -ct.

  • I can also confirm that on linux using OpenEdge 11.7 with -ct 1 you get an almost instant error if the database you connect client/server is not running

    Without -ct it takes almost a minute (50+ secs)

  • -ct 1 should go in your .pf file in the $DLC directory.

  • Sounds like a great candidate for a new and improved default value.

    --
    Tom Bascom
    tom@wss.com

  • -ct in startup.pf does not work, you need to specify it for each database connection seperately

  • > On Jul 4, 2019, at 10:02 AM, cverbiest wrote:

    >

    > -ct in startup.pf does not work, you need to specify it for each database connection seperately

    >

    >

    >

    of course it doesn't (since it is a database connection parameter). i was a fool to suggest it.

  • "of course it doesn't (since it is a database connection parameter). i was a fool to suggest it."

    C'mon Gus, you were just testing to see if people still read your posts ;-)