RE: shmget EEXIST 17 : shared memory already exists
Hello,
Thanks for these informations.
I checked the cron, but nothing is suspect. The only thing is that a backup
start each day at 3.00 AM.
Maybe the system activity is more intense during backup ?? (and Informix
claims more memory ??)
I know checkpoints occurs too often, but it was the easiest way to reduce
their time (I know it's shame :-((()
The checkpoint at 3:21:13 is strange, but it's maybe because we have two
server as 'cluster' and their time is not synchronised ??
I changed these two SHM parameters, but with no luck :
SHMVIRTSIZE 128000
SHMADD 32000
The server crashes about once a week ... we'll continue to investigate.
Thanks
Ben
> -----Original Message-----
> From: Art S. Kagel [SMTP:kagel@bloomberg.net]
> Sent: Monday, April 17, 2000 6:37 PM
> To: informix-list@iiug.org
> Subject: Re: shmget EEXIST 17 : shared memory already exists
>
> Looks like some cron is trying to start the engine but it is already
> online.
> That
> checkpoint shown at 3:21:13 is not being generated by the engine that
> failed to
> allocate its resident segment! (BTW why are you checkpointing every 2
> minutes?
> Most of us checkpoint no more often than once every 15 minutes many as
> seldom
> as once an hour!) Is it possible that someone has a cron that is trying
> to
> restart the
> engine every night which if it gets an error to oninit then performs an
> onmode
> -ky> which is really why the engine is shutting down at 3:21:14.
>
> When we were using Oracle for a project for a while the Oracle techs
> installed a
>
> cron that would try to restart their engine every two hours "Just to be
> sure
> it's
> always online!" Odd how I've never gotten advice to do something like
> that from
>
> an Informix tech. Anyway it sounds a lot like that kind of a paranoia
> cron.
>
> Henry Benoit wrote:
>
> > Hello,
> >
> > We use : Informix Dynamic Server Version 7.31.UC2A
> > on a Siemens box : ReliantUNIX-Y 5.45 A1023 RM600 6/3072 R10000
> >
> > About twice a month, our db server crashes with this error :
> > shmget: [EEXIST][17]: key 52564801: shared memory already exists
> >
> > -----
> > 03:15:52 Checkpoint Completed: duration was 0 seconds.
> >
> > 03:17:53 Checkpoint Completed: duration was 0 seconds.
> >
> > 03:19:54 Checkpoint Completed: duration was 0 seconds.
> >
> > 03:21:08 shmget: [EEXIST][17]: key 52564801: shared memory already> exists
> >
> > 03:21:08 mt_shm_init: can't create resident segment> >
> >
> >
> > 03:21:13 Checkpoint Completed: duration was 0 seconds.
> >
> > 03:21:14 Informix Dynamic Server Stopped.> >
> > -----
> >
> > I think that something is wrong with shared memory configuration,
> > but I can't see which parameter. (Our server have 3 GB RAM)
> >
> > >onstat -c | grep SHM> > SHMBASE 0x20000000 # Shared memory base address> >
> > #SHMVIRTSIZE 8000 # initial virtual shared memory segment
> size
> >
> > #SHMVIRTSIZE 245760 # initial virtual shared memory segment
> > size
> > SHMVIRTSIZE 16000
> >
> > SHMADD 8192 # Size of new shared memory segments
> > (Kbytes)
> > SHMTOTAL 0 # Total shared memory (Kbytes).
> 0=>unlimited> >
> > Is there any way to avoid these crashes at night ?
> > Can someone help me to better configure the shared memory ?
> >
> > Thanks in advance
> >
> > Ben