Re: INFORMIX CHECKPOINT REQUEST AND... FREEZE!!!

Tech suppor are evaluating the situation... but I really don't trust them. I must wait... Thanks a lot. Pablo Neil Truby ha escrito: > "Pablo" <micasilla@gmail.com> wrote in message > news:1150166383.416336.13170@y43g2000cwc.googlegroups.com... > > Hi! > > IBM Informix Dynamic Server Version 9.40.UC6 on RedHat Release 9 > > > > Sinse last week, 3 o 4 time at day, its freeze, > > i mean: > > On-Line (CKPT REQ) -- Up 01:32:43 -- 636720 Kbytes Blocked:CKPT > > > > And no online.log mesage, no console mensage, no error mensage: > > nothing, just halt... > > And continue reciving conections that can't resolve.... so, the > > userthreas numbers make bigger and bigger... until I have to kill the > > oninit (because I cant execute onmode -ky), and restart the > > server/engenie (oninit -v) > > > > After the restart... its begin a "fast recovery" until it puts > > ONLINE... > > > > Any solution? > > THANKS!!! > > > > -Pablo (Buenos Aires) > > > > I saw a problem in a patched version of IDS 10.0xC3 whereby a user thread > could just hold on of the cpu vps forever. However that became apparent > almost immediately whereas you imply that 9.40UC6 was working fine and now > isn't. > > Tech Support call?

interesting you don't "trust" tech support. assuming you have support with IBM and assuming you can afford some downtime and assuming you have explained this is happening 3 or 4 times a day and is seriously affecting your business I would hope that IBM would dial in and have a look at what is going on at the time the problem occurs. Its unlikely they will get much information when the problem is not occuring. if you are not satisfied with the service they are providing it usually helps to ask to speak with a manager or ask for the managers email address so you can contact them electronically

Pablo wrote: > Tech suppor are evaluating the situation... > but I really don't trust them. You should trust them. > I must wait... > You do not need to wait. Next time it happens use onstat -o to take a shared memory dump then sent it to IBM. They should be able to tell from that what the problem is.