Long checkpoints

Hi, all. Environment: IDS 9.21, HP UX 11, OLTP application. Informix stores data on raw devices on disk array. Unfortunatley, this array is used by other applications also. Guess, this is reason, why sometimes checkpoint time jumps to 20 seconds, which is not acceptable. Below is current config file. With LRU_MIN_DIRTY and LRU_MAX_DIRTY set to 0 and 1 and all other parameters set as they are, checkpoints should happen often and be very short, as far as I understand. But with all this set, checkpoints happen at CKPTINTVL, every 3 minutes, so I think that AMOUNT of data being written is not what causes the long checkpoint. Am I right, assuming that this is not Informix related problem and we should configure disk array not tune Informix parameters? Below is also sar -d, how about it? Thanks! Andris # Root Dbspace Configuration ROOTNAME rootdb # Root dbspace nameROOTPATH /cortex/dev/rootdb # Path for device containing root dbspace ROOTOFFSET 0 # Offset of root dbspace into device (Kbytes) ROOTSIZE 100000 # Size of root dbspace (Kbytes) # Disk Mirroring Configuration Parameters MIRROR 0 # Mirroring flag (Yes = 1, No = 0) MIRRORPATH # Path for device containing mirroredroot MIRROROFFSET 0 # Offset into mirrored device (Kbytes) # Physical Log Configuration PHYSDBS plog # Location (dbspace) of physical log PHYSFILE 12000 # Physical log file size (Kbytes) # Logical Log Configuration LOGFILES 876 # Number of logical log files LOGSIZE 4000 # Logical log size (Kbytes) # Diagnostics MSGPATH /ctxtools/informix/online2000.log # System message log file path CONSOLE /dev/console # System console message path ALARMPROGRAM /ctxtools/informix/etc/alarm_prog.sh # Alarm programpath TBLSPACE_STATS 0 # Maintain tblspace statistics # System Archive Tape Device TAPEDEV /dev/tape # Tape device path TAPEBLK 512 # Tape block size (Kbytes) TAPESIZE 100000000 # Maximum amount of data to put on tape (Kbytes) # Log Archive Tape Device LTAPEDEV /cortex/dev/tapen # Log tape device path LTAPEBLK 512 # Log tape block size (Kbytes) LTAPESIZE 100000000 # Max amount of data to put on log tape (Kbytes) # System Configuration SERVERNUM 0 # Unique id corresponding to a OnLineinstance DBSERVERNAME cortex # Name of default database server DBSERVERALIASES cortex_tcp # List of alternate dbservernames NETTYPE ipcshm,2,320,CPU # Configure poll thread(s) fornettype #NETTYPE soctcp,1,270,NET # Configure poll thread(s) for nettype DEADLOCK_TIMEOUT 60 # Max time to wait of lock indistributed env. RESIDENT 0 # 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 vpsto one NOAGE 1 # Process aging AFF_SPROC 0 # Affinity start processor AFF_NPROCS 0 # Affinity number of processors # Shared Memory Parameters LOCKS 500000 # Maximum number of locks BUFFERS 393216 # Maximum number of shared buffers NUMAIOVPS 20 # Number of IO vps PHYSBUFF 32 # Physical log buffer size (Kbytes) LOGBUFF 32 # Logical log buffer size (Kbytes)LOGSMAX 4000 # Maximum number of logical log files CLEANERS 128 # Number of buffer cleaner processes SHMBASE 0x0 # Shared memory base address SHMVIRTSIZE 158940 # initial virtual shared memorysegment size SHMADD 32768 # Size of new shared memory segments (Kbytes) SHMTOTAL 0 # Total shared memory (Kbytes). 0=>unlimited CKPTINTVL 180 # Check point interval (in sec) LRUS 128 # Number of LRU queues LRU_MAX_DIRTY 1 # LRU percent dirty begin cleaninglimit LRU_MIN_DIRTY 0 # 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 32 # Stack size (Kbytes) DBSPACETEMP # Default temp dbspaces --------------------------------------------------------------------------- 12:55:00 device %busy avque r+w/s blks/s avwait avserv 12:55:03 c2t2d0 4.59 0.59 6 46 5.20 14.02 c1t2d0 3.79 0.50 5 42 4.84 10.64 c23t2d1 96.41 8.77 248 2191 80.09 34.97 12:55:08 c2t2d0 6.96 0.50 8 45 5.09 9.74 c1t2d0 3.98 0.50 5 29 4.55 8.06 c23t2d1 98.21 1.91 60 1705 38.82 92.98 12:55:13 c2t2d0 5.65 0.50 6 40 5.57 13.62 c1t2d0 3.83 0.50 4 35 5.40 9.50 c23t2d1 97.18 0.94 79 1782 38.33 77.97 12:55:18 c2t2d0 3.79 0.50 4 16 4.90 10.63 c1t2d0 2.99 0.50 3 14 4.98 9.51 c23t2d1 59.68 0.84 78 1864 14.37 29.00 12:55:23 c2t2d0 5.19 0.50 7 74 4.87 15.57 c1t2d0 4.59 0.50 6 68 4.62 14.30 c23t2d1 97.80 3.17 73 1931 26.42 61.09 12:55:28 c2t2d0 3.41 0.50 4 43 6.38 10.72 c1t2d0 2.40 0.50 3 39 6.71 8.99 c23t2d1 96.39 0.84 61 1630 18.46 91.71 12:55:33 c2t2d0 3.01 0.50 3 19 5.44 10.64 c1t2d0 2.20 0.50 2 15 5.17 9.75 c23t2d1 55.51 0.60 57 968 8.16 30.87 12:55:38 c2t2d0 4.77 0.50 5 20 5.87 10.33 c1t2d0 3.18 0.50 4 16 6.24 7.98 c23t2d1 99.40 6.14 59 2128 98.71 120.28 12:55:43 c2t2d0 3.62 0.50 5 48 5.36 9.72 c1t2d0 3.02 0.50 4 42 5.88 9.70 c23t2d1 84.71 0.69 97 1891 13.22 36.93 12:55:48 c2t2d0 2.20 0.50 3 22 4.98 9.78 c1t2d0 1.60 0.50 2 19 5.33 8.88 c23t2d1 13.20 0.58 52 1473 5.25 18.20 12:55:53 c2t2d0 2.19 0.50 3 20 4.81 9.50 c1t2d0 1.99 0.50 2 20 4.79 8.29 c23t2d1 66.53 0.91 82 1648

c23t2d1 96.41 8.77 248 2191 80.09 34.97 Is the killer line, this disk is being flattened, 20% is typically deemed to be very busy, access time to 'standard' disks should be 10-20 ms, NVRAM backed arrays can run at <1ms, wait times should be 0 or < 5ms Cheers Paul Andris wrote: > > Hi, all. > > Environment: > > IDS 9.21, HP UX 11, OLTP application. > > Informix stores data on raw devices on disk array. Unfortunatley, this > array is > used by other applications also. Guess, this is reason, why sometimes > checkpoint time jumps to 20 seconds, which is not acceptable. Below is > current config file. With LRU_MIN_DIRTY and LRU_MAX_DIRTY set to 0 and > 1 and all other parameters set as they are, checkpoints should happen > often and be very short, as far as I understand. But with all this > set, checkpoints happen at CKPTINTVL, every 3 minutes, so I think that > AMOUNT of data being written is not what causes the long checkpoint. > Am I right, assuming that this is not Informix related problem and we > should configure disk array not tune Informix parameters? > [cutting] > 12:55:00 device %busy avque r+w/s blks/s avwait avserv > 12:55:03 c2t2d0 4.59 0.59 6 46 5.20 14.02 > c1t2d0 3.79 0.50 5 42 4.84 10.64 > c23t2d1 96.41 8.77 248 2191 80.09 34.97 > 12:55:08 c2t2d0 6.96 0.50 8 45 5.09 9.74 > c1t2d0 3.98 0.50 5 29 4.55 8.06 > c23t2d1 98.21 1.91 60 1705 38.82 92.98 > 12:55:13 c2t2d0 5.65 0.50 6 40 5.57 13.62 > c1t2d0 3.83 0.50 4 35 5.40 9.50 > c23t2d1 97.18 0.94 79 1782 38.33 77.97 > 12:55:18 c2t2d0 3.79 0.50 4 16 4.90 10.63 [cutting] -- Paul Watson # Oninit Ltd # Growing old is mandatory Tel: +44 1436 672201 # Growing up is optional Fax: +44 1436 678693 # Mob: +44 7818 003457 # www.oninit.com #