ONCONFIG for 9.40 offered for comment

9.40 FC3W2 on HP-UX 11.11 The server has 6 cpus and 6G of RAM. 3 other, small, IDS 9.40 instances run on the server, and one small 7.31. kaio is enabled. Just put 9.40 live, and it's running about 5-10% more slowly on some key processes. This reflects what happened in testing. In particular we seem to be having fun and games with the new BTscanners. I thought I'd post the ONCONFIG for comment - any and all gratefully received! thanks Neil P.S. We know we'll have to increase SHMVIRTSIZE, as full-throttle BT scanning breaks out more segments ... #************************************************************************** # # INFORMIX SOFTWARE, INC. # # Title: Onconfig for instance 1 # Description: INFORMIX-OnLine Configuration Parameters # #************************************************************************** # Root Dbspace Configuration ROOTNAME rootdbs # Root dbspace nameROOTPATH /opt/informixV7.13/dbspaces1/rootdbs # Note from Neil - before my time!!!!! ROOTOFFSET 0 # Offset of root dbspace into device (Kbytes) ROOTSIZE 131072 # Size of root dbspace (Kbytes) # Disk Mirroring Configuration Parameters MIRROROFFSET 0 # Offset into mirrored device (Kbytes) # Physical Log Configuration PHYSDBS physlog # Location (dbspace) of physical log PHYSFILE 80000 # Physical log file size (Kbytes) # Logical Log Configuration LOGFILES 50 # Number of logical log files LOGSIZE 10240 # Logical log size (Kbytes) # Diagnostics MSGPATH /opt/informix/online_1.log # System message log file path CONSOLE /opt/informix/online_1.con # System console message path ALARMPROGRAM /opt/informix/9.40/etc/log_full.sh # Alarm program path # System Archive Tape Device TAPEDEV /opt/informixV7.13/dbspaces1/tapedev TAPEDEV /dev/null TAPEBLK 64 # Tape block size (Kbytes) TAPESIZE 40960000 # Maximum amount of data to put on tape (Kbytes) # Log Archive Tape Device LTAPEDEV /opt/informixV7.13/dbspaces1/ltapedev # Log tape device path LTAPEBLK 16 # Log tape block size (Kbytes) LTAPESIZE 8192000 # Max amount of data to put on log tape (Kbytes) # Optical STAGEBLOB # INFORMIX-OnLine/Optical staging area # System Configuration SERVERNUM 1 # Unique id corresponding to a OnLineinstance DBSERVERNAME lawson_live_shm # Name of default database server DBSERVERALIASES lawson_live,ifx_tcp_2,cats,bo_data,datastore,lawson,lawson_live_soc DEADLOCK_TIMEOUT 120 # Max time to wait of lock in distributed env. RESIDENT 1 # Forced residency flag (Yes = 1, No = 0) MULTIPROCESSOR 1 # 0 for single-processor, 1 formulti-processor NUMCPUVPS 3 # Number of user (cpu) vps SINGLE_CPU_VP 0 # If non-zero, limit number of cpu vps toone NOAGE 1 # Process aging AFF_SPROC 0 # Affinity start processor AFF_NPROCS 0 # Affinity number of processors # Shared Memory Parameters LOCKS 2000000 # Maximum number of locks BUFFERS 375000 # Maximum number of shared buffers NUMAIOVPS 3 # Number of IO vps#NUMAIOVPS 127 # Changed with upgrade to 9.40 as usingn KAIO instead PHYSBUFF 256 # Physical log buffer size (Kbytes) LOGBUFF 128 # Logical log buffer size (Kbytes)LOGSMAX 150 # Maximum number of logical log files CLEANERS 65 # Number of buffer cleaner processes SHMBASE 0x0 # Shared memory base address SHMVIRTSIZE 320768 # initial virtual shared memory segment size SHMADD 16384 # Size of new shared memory segments (Kbytes) SHMTOTAL 0 # Total shared memory (Kbytes). 0=>unlimited CKPTINTVL 900 # Check point interval (in sec) LRUS 65 # Number of LRU queues LRU_MAX_DIRTY 4 # LRU percent dirty begin cleaning limit LRU_MIN_DIRTY 2 # LRU percent dirty end cleaning limit LTXHWM 50 # Long transaction high water markpercentage LTXEHWM 60 # Long transaction high water mark (exclusive) TXTIMEOUT 0x12c # Transaction timeout (in sec) STACKSIZE 128 # Stack size (Kbytes) # System Page Size # BUFFSIZE - OnLine no longer supports this configuration parameter. # To determine the page size used by OnLine on your platform # see the last line of output from the command, 'onstat -b'. # Recovery Variables # OFF_RECVRY_THREADS: # Number of parallel worker threads during fast recovery or an offline restore. # ON_RECVRY_THREADS: # Number of parallel worker threads during an online restore. OFF_RECVRY_THREADS 10 # Default number of offline workerthreads ON_RECVRY_THREADS 1 # Default number of online worker threads # Data Replication Variables # DRAUTO: 0 manual, 1 retain type, 2 reverse type DRINTERVAL 30 # DR max time between DR buffer flushes (in sec) DRTIMEOUT 30 # DR network timeout (in sec) DRLOSTFOUND /dev/null # DR lost+found file path # Read Ahead Variables RA_PAGES 32 # Number of pages to attempt to read ahead RA_THRESHOLD 16 # Number of pages left before next group # DBSPACETEMP: # OnLine equivalent of DBTEMP for SE. This is the list of dbspaces # that the OnLine SQL Engine will use to create temp tables etc. # If specified it must be a colon separated list of dbspaces that exist # when the OnLine system is brought online. If not specified, or if # all dbspaces specified are invalid, various ad hoc queries will create # temporary files in /tmp instead. DBSPACETEMP dbspacetemp00,dbspacetemp01,dbspacetemp02 # Default temp dbspaces # DUMP*: # The following parameters control the type of diagnostics information which # is preserved when an unanticipated error condition (assertion failure) occurs # during OnLine operations. # For DUMPSHMEM, DUMPGCORE and DUMPCORE 1 means Yes, 0 means No. DUMPDIR /var/tmp # Preserve diagnostics in this directory DUMPSHMEM 0 # Dump a copy of shared memory DUMPGCORE 0 # Dump a core image using 'gcore' DUMPCORE 0 # Dump a core image (Warning:this aborts OnLine) DUMPCNT 1 # Number of shared memory or gcore dumps for # a single user's session # ADT* # The following parameters control the type and level of secure auditing # present in the OnLine system. By default, ADTMODE is 0 and auditin

Neil Truby wrote: > 9.40 FC3W2 on HP-UX 11.11 > The server has 6 cpus and 6G of RAM. > 3 other, small, IDS 9.40 instances run on the server, and one small > 7.31. kaio is enabled. > > Just put 9.40 live, and it's running about 5-10% more slowly on some > key processes. This reflects what happened in testing. In particular > we seem to be having fun and games with the new BTscanners. I > thought I'd post the ONCONFIG for comment - any and all gratefully > received! Are the small instances treated 2nd-class? ie they don't get NOAGE or affinity? I would give all the good stuff to the primary engine and let the others suffer if they are only little test engines. It all looks groovy, except: > TAPEBLK 64 # Tape block size (Kbytes) > LTAPEBLK 16 # Log tape block size HP works really well with larger blocks for this - ie 256 or even 512. > NUMCPUVPS 3 # Number of user (cpu) vps Give the big engine more CPU's? I guess this config is unchanged since the previous version? > NOAGE 1 # Process aging > AFF_SPROC 0 # Affinity start processor > AFF_NPROCS 0 # Affinity number of processors AFF_ works well on HP - try 1,3 and make sure the other engines have NOAGE=0 and AFF_ =0 > RA_PAGES 32 > RA_THRESHOLD 16 # Number of pages left before> next group What it says, so the read-ahead might not keep up. I go with 32/28 or 32/24 ie something which sends it on it's way a bit earlier. > NETTYPE soctcp,2,75,CPU # Override sqlhosts nettype parameters > NETTYPE ipcshm,1,25,NET # Override sqlhosts nettype parameters Why are these "backwards"? Swap the CPU and NET. I'd be inclined to go 1,150 and 3,25 respectively, although your numbers are quite small so it's probably not a big deal either way. > TBLSPACE_STATS 1 This is said to cause the engine to burn more cycles than it needs to. Apart from that, I haven't personally met 9.40 although I'm looking at it now. What else has changed?

"Andrew Hamm" <ahamm@mail.com> wrote in message news:2ja9l4Fuctd6U1@uni-berlin.de... > Neil Truby wrote: > > 9.40 FC3W2 on HP-UX 11.11 > > The server has 6 cpus and 6G of RAM. > > 3 other, small, IDS 9.40 instances run on the server, and one small > > 7.31. kaio is enabled. > > > > Just put 9.40 live, and it's running about 5-10% more slowly on some > > key processes. This reflects what happened in testing. In particular > > we seem to be having fun and games with the new BTscanners. I > > thought I'd post the ONCONFIG for comment - any and all gratefully > > received! > > Are the small instances treated 2nd-class? ie they don't get NOAGE or > affinity? I would give all the good stuff to the primary engine and let the > others suffer if they are only little test engines. All the little fellas have 2 virtual cpus allocated - although small they are important, and we've had problems before with a single cpu vp being locked out for long periods with a runaway process - but NOAGE set to 0 and no processor affinity anyway. > It all looks groovy, except: > > > TAPEBLK 64 # Tape block size (Kbytes) > > LTAPEBLK 16 # Log tape block size> > HP works really well with larger blocks for this - ie 256 or even 512. Fair point but in fact this database is rarely archived via Informix utilities - we use sexy state-of-the-art SAN integration and external restores - but when it is it's done via onbar and a storage manager anyway. > > NUMCPUVPS 3 # Number of user (cpu) vps> > Give the big engine more CPU's? I guess this config is unchanged since the > previous version? Yes, it is unchanged from 9.21, although we have enabld kaio with 9.40, having had it disabled with 9.21. I've tried dynamically adding some but it makes little apparent difference. > > NOAGE 1 # Process aging > > AFF_SPROC 0 # Affinity start processor > > AFF_NPROCS 0 # Affinity number of processors> > AFF_ works well on HP - try 1,3 and make sure the other engines have NOAGE=0 > and AFF_ =0 > > RA_PAGES 32 > > RA_THRESHOLD 16 # Number of pages left before> > next group > > What it says, so the read-ahead might not keep up. I go with 32/28 or 32/24 > ie something which sends it on it's way a bit earlier. > > NETTYPE soctcp,2,75,CPU # Override sqlhosts nettype parameters > > NETTYPE ipcshm,1,25,NET # Override sqlhosts nettype parameters> > Why are these "backwards"? Swap the CPU and NET. I'd be inclined to go 1,150 > and 3,25 respectively, although your numbers are quite small so it's > probably not a big deal either way. > > TBLSPACE_STATS 1> > This is said to cause the engine to burn more cycles than it needs to. > Apart from that, I haven't personally met 9.40 although I'm looking at it > now. > > What else has changed? The BT scanner is the big change. There's some stuff about experiences on our website (www.ardenta.com). cheers Neil

On Wed, 16 Jun 2004 01:01:12 -0400, Neil Truby wrote: Neil, I'd recommend the following: > 9.40 FC3W2 on HP-UX 11.11 > The server has 6 cpus and 6G of RAM. > 3 other, small, IDS 9.40 instances run on the server, and one small 7.31. > kaio is enabled. > > Just put 9.40 live, and it's running about 5-10% more slowly on some key > processes. This reflects what happened in testing. In particular we seem to > be having fun and games with the new BTscanners. I thought I'd post the > ONCONFIG for comment - any and all gratefully received! <SNIP> > NETTYPE soctcp,2,75,CPU # Override sqlhosts nettype parameters > NETTYPE ipcshm,1,25,NET # Override sqlhosts nettype parameters TCP connections should be handled by the NET VPs and shared memory connections by the CPU VP to minimize system overhead and maximize responsiveness (I know different from the manuals prior to 9.4). Also the listeners for shm connections should run in every CPU VP. So: NETTYPE soctcp,2,75,NET NETTYPE ipcshm,3,25,CPU I do not see any entry for NUMAIOVPS or the equivalent VPCLASS entry (preferred in 9.4, NUMAIOVPS is going away as is NUMCPUVPS), even with KAIO enabled, you need more than the default number of AIO VPs. I'd start with 6 and monitor onstat -g iov, indeed if you look now you will likely see only 2 or 3 AIO VPs and the io/wup for all will be >= 1.0 which means you need more of them. Believe it or not the message log activity which these VPS handle even with KAIO enabled, can hurt server performance if it backs up. What do the BR, BTR, & RAU ratios look like? Art S. Kagel

"Art S. Kagel" <kagel@bloomberg.net> wrote in message news:pan.2004.06.16.11.49.42.178367.9956@bloomberg.net... > On Wed, 16 Jun 2004 01:01:12 -0400, Neil Truby wrote: > > Neil, I'd recommend the following: > I do not see any entry for NUMAIOVPS or the equivalent VPCLASS entry > (preferred in 9.4, NUMAIOVPS is going away as is NUMCPUVPS), even with KAIO > enabled, you need more than the default number of AIO VPs. I'd start with 6 > and monitor onstat -g iov, indeed if you look now you will likely see only 2 > or 3 AIO VPs and the io/wup for all will be >= 1.0 which means you need more > of them. Believe it or not the message log activity which these VPS handle > even with KAIO enabled, can hurt server performance if it backs up. It was in there somewhere, set to 3. Here's the iov output (stats zeroed out a few hours ago): IBM Informix Dynamic Server Version 9.40.FC3W2 -- On-Line -- Up 4 days 03:59:52 -- 1507604 Kbytes AIO I/O vps:class/vp s io/s totalops dskread dskwrite dskcopy wakeups io/wup errors kio 0 i 366.6 15518033 14544634 973399 0 25112136 0.6 0 kio 1 i 185.6 7856279 7048573 807706 0 12975527 0.6 0 kio 2 s 188.3 7969620 7142574 827046 0 13188352 0.6 0 kio 3 s 0.0 0 0 0 0 0 0.0 0 kio 4 s 0.0 0 0 0 0 0 0.0 0 kio 5 s 0.0 0 0 0 0 0 0.0 0 kio 6 s 0.0 0 0 0 0 0 0.0 0 msc 0 i 0.5 20009 0 0 0 18763 1.1 0 aio 0 i 0.2 7041 5390 1429 0 4637 1.5 0 aio 1 i 0.1 3379 1901 1478 0 980 3.4 0 aio 2 i 0.1 3117 1697 1420 0 866 3.6 0 pio 0 i 0.0 0 0 0 0 0 0.0 0 lio 0 i 0.0 0 0 0 0 0 0.0 0 > What do the BR, BTR, & RAU ratios look like? This is what I got from your "ratios" proc: $ dbaccess sysmaster << ! > execute procedure ratios () ;> ! Database selected. (expression) 1.50 (expression) 96.91 (expression) 16.57 (expression) 16.57 (expression) 12:05:50 (expression) 2004-06-16 06:23:09 cheers! Neil