Dynaverse.net

Taldrenites => Dynaverse II Experiences => Topic started by: KAT Chuut-Ritt on December 02, 2004, 09:44:53 am

Title: Bonk we might have broke it
Post by: KAT Chuut-Ritt on December 02, 2004, 09:44:53 am

The Stress Test server that is.  Seems we got 4 accounts ghosted and another one who's model wont load in mission.

Title: Re: Bonk we might have broke it
Post by: KDS-KYTARH on December 02, 2004, 09:58:45 am

Noooooooooooooooooooooooo!!!!!!!!!!!!!!!!!!!!!!! I just got my Romulan Condor last night!!


Bastiches! >:(

Title: Re: Bonk we might have broke it
Post by: Bonk on December 02, 2004, 10:29:22 am

I'll check it out, maybe it finally needs a reboot.

The thread count has more than doubled since this morning, and appears to be rising still (previously was in the range of 18-24, now at 64-88 range...)

Title: Re: Bonk we might have broke it
Post by: Bonk on December 02, 2004, 10:52:17 am

Just logged in, ran a mission OK, flipped a hex. Looks like some ghosts on there tho, I'm guessing network hiccup, when exactly did the trouble start? (I backed up the db at 11 AM AST while live - which might have caused troubles, the sql db is up to 1.2 MB now [battle logging and created ai filling it])

The thread count continues to steadily rise... must be due to the ghosted accounts... I'll reboot it now so its fresh for tonight, I'm hoping to get 25 on still to see what happens. (4 days without a reboot - not bad I guess)

Seemed to be about 78 threads that would not close when I tried to shutdown the kit using its menu, its thread and memory use (~63MB mem) only stopped when I closed the window. Its thread count and memory use are back down to normal after a reboot (~14 threads ~ 18MB mem). If I see the same process happens faster with 25 players on in a short time (over and above threads normally created by active players), as opposed to with just a few on over 4 days, then I can take a hard look at the kit to find the cause. (I do expect some mild improvement with this build however, if so I'll continue my edits along the same lines)

The "normal" thread count hovers around the number of tables in the SQL db with no players on, this event and observations leads me to believe I'm on the right track, I can explain better later... its adding up... (I'd like to study the flatfile in the same way for comparison sometime)

Title: Re: Bonk we might have broke it
Post by: benbean on December 02, 2004, 11:20:15 am

It's good to see your getting somewhere with this Bonk.

Title: Re: Bonk we might have broke it
Post by: RazalYllib on December 02, 2004, 06:10:43 pm

More Karma for Bonk for the dedication to the task...