Very large instances
IDS 11.70FC5 on RHEL 64-bit: We're running OK with about 110GB buggers and 90G SHMVIRTSIZE. We tried allocating some of an extra 256GB of RAM recently added to the server. We increased the huge pages by 70GB (sysctl -w vm.nr_hugepages=145000), and tried using 60GB of that for informix shm and buffers: 10GB extra for virtual 25GB extra for 8k and 16k bufferpools But it wouldnt work. SHMTOTAL is 0 and kernel.shmmax is 4Tb. Messages below - basically "Contiguous shared memory segment allocation failed" is Bad News and your oncommands are going to fail sporadically, so the instance has to be re-started until you don't get them. We'll raise a call with IBM when they're open, but I just wondered what other very large instances are out there? Thanks Neil ========================================================== Fri Nov 22 11:12:44 2013 11:12:44 Event alarms enabled. ALARMPROG = '/opt/informix/INST1/scripts/alarmprogram.sh' 11:12:44 Booting Language <c> from module <> 11:12:44 Loading Module <CNULL> 11:12:44 Booting Language <builtin> from module <> 11:12:44 Loading Module <BUILTINNULL> 11:12:55 DR: DRAUTO is 0 (Off) 11:12:55 DR: ENCRYPT_HDR is 0 (HDR encryption Disabled) 11:12:55 shmat: [22]: operating system error 11:12:55 Client could not attach server shared memory segment, use IFX_XFER_SHMBASE. 11:12:55 (2) shm creation of shmem segment failed 11:15:45 IBM Informix Dynamic Server Started. 11:15:45 Contiguous shared memory segment allocation failed at 0x25e0edd000. Allocation successful at 0x2aaaaee2e000. Check SHMBASE is consistent with the value in $INFORMIXDIR/etc/onconfig.std. If you are using the correct SHMBASE value in your ONCONFIG file, then consider this message informational only.
FYI, no direct help here, but v12.10.xC2 no longer has to keep the buffer pool in the resident segment, it has been moved into a virtual segment of its own and can grow beyond that single segment so the requirement for the buffer pool to be contiguous in memory is gone. Art Art S. Kagel, Principal Consultant Advanced DataTools (www.advancedatatools.com) Blog: http://informix-myview.blogspot.com/ Disclaimer: Please keep in mind that my own opinions are my own opinions and do not reflect on my employer, Advanced DataTools, the IIUG, nor any other organization with which I am associated either explicitly, implicitly, or by inference. Neither do those opinions reflect those of other individuals affiliated with any entity with which I am affiliated nor those of the entities themselves. On Sun, Nov 24, 2013 at 4:24 PM, NEIL TRUBY <neil.truby@ardenta.com> wrote: > IDS 11.70FC5 on RHEL 64-bit: > > We're running OK with about 110GB buggers and 90G SHMVIRTSIZE. > > We tried allocating some of an extra 256GB of RAM recently added to the > server. > We increased the huge pages by 70GB (sysctl -w vm.nr_hugepages=145000), and > tried using 60GB of that for informix shm and buffers: > > 10GB extra for virtual > 25GB extra for 8k and 16k bufferpools > > But it wouldnt work. > > SHMTOTAL is 0 and kernel.shmmax is 4Tb. > > Messages below - basically "Contiguous shared memory segment allocation > failed" is Bad News and your oncommands are going to fail sporadically, so > the > instance has to be re-started until you don't get them. > > We'll raise a call with IBM when they're open, but I just wondered what > other > very large instances are out there? > > Thanks > Neil > ========================================================== > Fri Nov 22 11:12:44 2013 > > 11:12:44 Event alarms enabled. ALARMPROG = > '/opt/informix/INST1/scripts/alarmprogram.sh' > 11:12:44 Booting Language <c> from module <> > 11:12:44 Loading Module <CNULL> > 11:12:44 Booting Language <builtin> from module <> > 11:12:44 Loading Module <BUILTINNULL> > 11:12:55 DR: DRAUTO is 0 (Off) > 11:12:55 DR: ENCRYPT_HDR is 0 (HDR encryption Disabled) > 11:12:55 shmat: [22]: operating system error > 11:12:55 Client could not attach server shared memory segment, use > IFX_XFER_SHMBASE. > 11:12:55 (2) shm creation of shmem segment failed > 11:15:45 IBM Informix Dynamic Server Started. > 11:15:45 Contiguous shared memory segment allocation failed at > 0x25e0edd000. > Allocation successful at 0x2aaaaee2e000. > Check SHMBASE is consistent with the value in > $INFORMIXDIR/etc/onconfig.std. > If you are using the correct SHMBASE value in your ONCONFIG file, then > consider this message informational only. > > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > --047d7b3a84c292487d04ebf39d3b
Hi Neil, Have you sorted this yet? If so, what was the solution? Some ideas: What is kernel.shmall set to? 145000 is a lot of huge pages (283 Gb assuming Linux 2 Mb huge page size). Is this the only instance on the box and are you using all of this for it? Is RESIDENT set to 2? (Adding extra memory segments into huge pages doesn't work on the Linux platform so this is the only sensible value.) Ben.
PMR 92115,019,866 refers btw