• Old bugbears

    From Spectre@21:3/105 to Somebody on Thursday, September 26, 2019 22:07:00
    One of them has come home to roost. Notably the time discrepancy problem. It used to be FD that always went walkabout. That system seems pretty stable, and the bbs nodes on the Veriton are wandering around. I don't think we really got to a reason why last time or come up with a solution.

    So from the top for a recap, buntu/virtualbox/MS-DOS time wandering around by ~5 mins in 2 hours or so. At the moment they won't stay close enough to have the BBS events coincide with FD events to suspend BBS action during mail tossing.

    All thoughts gratefully considered. I wonder is this is actually a DOS problem.....

    Spec


    --- --- SuperBBS v1.17-3 (Eval)
    * Origin: <Shoot'n the breeze on The Lower Planes> (21:3/105)
  • From Spectre@21:3/105 to Spectre on Thursday, September 26, 2019 22:48:00
    Bugbear addendum

    Short term for the moment, I've set mTCP's sntp client to run every cycle of the BBS batch file. However its accuracy doesn't seem to be better than a 20 minute window, up to 10mins behind or in front... which isn't that helpful in itself.

    Spec


    --- --- SuperBBS v1.17-3 (Eval)
    * Origin: <Shoot'n the breeze on The Lower Planes> (21:3/105)
  • From tallship@21:2/104 to Spectre on Friday, September 27, 2019 05:02:34
    On 26 Sep 2019, Spectre said the following...

    So from the top for a recap, buntu/virtualbox/MS-DOS time wandering
    around by ~5 mins in 2 hours or so. At the moment they won't stay close


    Arrggghh!!

    That's not good. You need to hunt that down, but my first concern would be whethr it was hardware related, and if so, what that might mean moving
    forward without replacing the hardware.

    But for the sake of keeping something going, a kludge could be a cronjob calling something like

    ntpdate ntp.ucsd.edu

    every few minutes.

    But something is bad wrong.

    --- Mystic BBS v1.12 A43 2019/03/02 (Linux/64)
    * Origin: Vger.Cloud - NOMAD Internetwork (21:2/104)
  • From Spectre@21:3/105 to tallship on Friday, September 27, 2019 23:46:00
    That's not good. You need to hunt that down, but my first concern
    would be whethr it was hardware related, and if so, what that might
    mean moving forward without replacing the hardware.

    Its not hardware related, system time stays pretty right, and it has been set to ntp check previously. Its actually the VM times that wander around. Each VM, they all run MS-DOS is presently running ntp whenever the BBS cycles through its batch file. Which is actually working better than I thought, but a long term solution would be nice.

    Spec


    --- --- SuperBBS v1.17-3 (Eval)
    * Origin: <Shoot'n the breeze on The Lower Planes> (21:3/105)
  • From Spectre@21:3/105 to Somebody on Thursday, September 26, 2019 22:07:00
    One of them has come home to roost. Notably the time discrepancy problem. It used to be FD that always went walkabout. That system seems pretty stable, and the bbs nodes on the Veriton are wandering around. I don't think we really got to a reason why last time or come up with a solution.

    So from the top for a recap, buntu/virtualbox/MS-DOS time wandering around by ~5 mins in 2 hours or so. At the moment they won't stay close enough to have the BBS events coincide with FD events to suspend BBS action during mail tossing.

    All thoughts gratefully considered. I wonder is this is actually a DOS problem.....

    Spec


    --- --- SuperBBS v1.17-3 (Eval)
    * Origin: <Shoot'n the breeze on The Lower Planes> (21:3/105)
  • From Spectre@21:3/105 to Spectre on Thursday, September 26, 2019 22:48:00
    Bugbear addendum

    Short term for the moment, I've set mTCP's sntp client to run every cycle of the BBS batch file. However its accuracy doesn't seem to be better than a 20 minute window, up to 10mins behind or in front... which isn't that helpful in itself.

    Spec


    --- --- SuperBBS v1.17-3 (Eval)
    * Origin: <Shoot'n the breeze on The Lower Planes> (21:3/105)
  • From Spectre@21:3/105 to tallship on Friday, September 27, 2019 23:46:00
    That's not good. You need to hunt that down, but my first concern
    would be whethr it was hardware related, and if so, what that might
    mean moving forward without replacing the hardware.

    Its not hardware related, system time stays pretty right, and it has been set to ntp check previously. Its actually the VM times that wander around. Each VM, they all run MS-DOS is presently running ntp whenever the BBS cycles through its batch file. Which is actually working better than I thought, but a long term solution would be nice.

    Spec


    --- --- SuperBBS v1.17-3 (Eval)
    * Origin: <Shoot'n the breeze on The Lower Planes> (21:3/105)
  • From Spectre@21:3/105 to Blue White on Thursday, October 10, 2019 02:27:00
    around by ~5 mins in 2 hours or so. At the moment they
    won't stay
    close enough to have the BBS events coincide with FD
    events to suspend
    BBS action during mail tossing.

    Is it wandering on the whole system, or just in the DOS VM?

    It only happens on the DOS VM's notably the one driving FrontDoor is the worst for some reason.

    Spec


    --- --- SuperBBS v1.17-3 (Eval)
    * Origin: <Shoot'n the breeze on The Lower Planes> (21:3/105)