Re: dynamically allocated new shared memory segment - What is this?

From: Gary Quiring <gquiring@netlabs.net> > >On Thu, 06 Apr 2000 08:02:56 BST, "Obnoxio The Clown" ><obnoxio@hotmail.com> wrote: > > >From: Gary Quiring <gquiring@netlabs.net> > >> > >>We had a pause in our Informix 7.3 system today around 5:00pm. The > >>system just 'froze' for almost a minute. Is there something that > >>needs to be tuned? This is the log during that period: > > > >Yeah, there's some code in the engine that detects that it's time to go >home > >and it just tries to complete all the work still outstanding by 5PM. > >Where's the option for 4:00pm!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! Sorry, it's hardwired in... > >This means the system was busy doing something that required a lot of > >virtual memory (like a big query or something? or it could be a bug) > > > >If it happens again, see if an onmode -F frees the virtual memory (check > >with onstat -g seg) -- if it does, it's just a temporary thing, if it > >doesn't, and you're sure that the thread causing the memory usage is >over, > >you may have a bug. > >Does this new shared memory allotment eventually get released, or do I >need to restart the engine? Here's the onstat -g seg for today. It should eventually be released. If it's not in use. One way to see if there is a problem is to disconnect all connections, and see if the virtual segments go away. >Informix Dynamic Server Version 7.30.UC2 -- On-Line -- Up 6 days >03:37:55 -- 8 >84736 Kbytes > >Segment Summary: >id key addr size ovhd class blkused >blkfree >1 1381386241 82000000 704643072 11604 R* 85557 459 >2 1381386242 ac000000 134217728 2640 V 5039 11345 >3 1381386243 b4000000 4194304 656 M 261 251 >4 1381386244 b4400000 4194304 656 M 259 253 >9 1381386245 b4800000 33554432 1104 V 185 3911 >10 1381386246 b6800000 33554432 1104 V 73 4023 >Total: - - 914358272 - - 91374 20242 All the V segments seem to have something in them. > >>16:58:20 Logical Log 23614 Complete. > >>17:00:17 Checkpoint Completed: duration was 32 seconds.> > > >That is one mean momma of a checkpoint -- 32 seconds? Did you do a batch > >load or mass update/delete? > >Nothing special going on, but the financials are NewERA and are not >written well. I see 1-6 long transactions being aborted in the log Well, there's an oxymoron for you. >daily. We have reported this to the software vendor, but no luck in >getting them to fix it. You are Donald Ducked, then, aren't you? :-/ ______________________________________________________ Get Your Private, Free Email at http://www.hotmail.com