Re: checkpoints taking as much as 25 secs
Try modifying your LRU's to something like 63. Make your page cleaners equal
to your LRU's value. I find I get considerable performance from my server
set this way. As well, checkpoints on high end machines running hundreds of
users using ~10GIG+ databases range from 1 -2 secs. on high peak times. but
sometimes get up to 5 or 6.
Try playing with these values here... to get the best result.
LRU=63
CLEANERS=63
Modify your LRU_MAX_DIRTY=2
LRU_MIN_DIRTY=1
HTH,
Lloyd AJ Wilson.
"Tam McLaughlin" <tamm@scotlegal.com> wrote in message
news:83sqbo$65b$1@neptunium.btinternet.com...
> The online log file shows that some checkpoints can take from 4 secs up to
> 25 secs
> but the average is prob 0 secs roughly every 5 mins. The long checkpoint
> times prob
> occur when some large progs are running e.g. cron jobs. I have not been on
> an IDS7
> admin course yet so not too sure what to look for. Our system is:
>
> IDS 7.30.UC2 on SCO UnixWare 7.0.1 with 3 pentium pro processors running
> RAID 5.
> We have approx 10 Gb of dbspace used and the largest table is ~ 3 Gb and
our
> app
> is mostly read only.
>
> I realise raid is not the best option and we may have a disk i/o
bottleneck.
> It may be that
> my question is too general and that it could take a long time to work out
> why we are getting
> these long checkpoints but if anyone has some clues where to start or what
> to check it may
> help me a great deal.
>
> Some info may help here:
>
> onstat -p : the bufwaits look high (i done an onstat -z a few days ago)but
> this may be normal
>
>
> Informix Dynamic Server Version 7.30.UC2 -- On-Line -- Up 21:14:51 --
> 114040 Kbytes
>
> Profile
> dskreads pagreads bufreads %cached dskwrits pagwrits bufwrits %cached
> 26169757 6148644 1242311687 97.89 140931 242136 5077195 97.22
>
> isamtot open start read write rewrite delete commit
> rollbk
> 162742135 3237700 9110352 121063233 2131212 225000 50794 40789
> 5439
>
> gp_read gp_write gp_rewrt gp_del gp_alloc gp_free gp_curs
> 0 0 0 0 0 0 0
>
> ovlock ovuserthread ovbuff usercpu syscpu numckpts flushes
> 0 0 0 37389.09 6435.06 214 508
>
> bufwaits lokwaits lockreqs deadlks dltouts ckpwaits compress seqscans
> 4750714 2 633141750 0 0 112 29328 384018
>
> ixda-RA idx-RA da-RA RA-pgsused lchwaits
> 15832893 12764 5910997 21753370 168421
>
>
>
> my onconfig file is:
>
> ROOTNAME rootdbs> ROOTPATH /dev/rdsk/rootdbs
> ROOTOFFSET 0
> ROOTSIZE 512000> MIRROR 0
> MIRRORPATH
> MIRROROFFSET 0> PHYSDBS rootdbs
> PHYSFILE 10000
> LOGFILES 21
> LOGSIZE 5000> MSGPATH /users/informix/pluto.log
> CONSOLE /users/informix/pluto_console.log
> ALARMPROGRAM /users/informix/etc/log_full.sh
> SYSALARMPROGRAM /users/informix/etc/evidence.sh
> TBLSPACE_STATS 1> STAGEBLOB
> SERVERNUM 0
> DBSERVERNAME pluto
> DBSERVERALIASES pluto_net
> NETTYPE ipcshm,1,64,CPU
> NETTYPE tlitcp,1,10,NET
> DEADLOCK_TIMEOUT 60
> RESIDENT 0
> MULTIPROCESSOR 1
> NUMCPUVPS 3
> SINGLE_CPU_VP 0
> NOAGE 0
> AFF_SPROC 0
> AFF_NPROCS 0
> LOCKS 20000
> BUFFERS 42000
> NUMAIOVPS 16
> PHYSBUFF 120
> LOGBUFF 40> LOGSMAX 60
> CLEANERS 6
> SHMBASE 0xa000000
> SHMVIRTSIZE 24000
> SHMADD 8192
> SHMTOTAL 0
> CKPTINTVL 300
> LRUS 8
> LRU_MAX_DIRTY 20
> LRU_MIN_DIRTY 15
> LTXHWM 45
> LTXEHWM 50
> TXTIMEOUT 0x12c
> STACKSIZE 32
> OFF_RECVRY_THREADS 10
> ON_RECVRY_THREADS 1
> DRAUTO 0
> DRINTERVAL 30
> DRTIMEOUT 30
> DRLOSTFOUND /users/informix/etc/dr.lostfound> CDR_LOGBUFFERS 2048
> CDR_EVALTHREADS 1,2
> CDR_DSLOCKWAIT 5
> CDR_QUEUEMEM 4096
> BAR_ACT_LOG /tmp/bar_act.log
> BAR_MAX_BACKUP 0
> BAR_RETRY 1
> BAR_NB_XPORT_COUNT 10
> BAR_XFER_BUF_SIZE 31> ISM_DATA_POOL ISMData
> ISM_LOG_POOL ISMLogs
> RA_PAGES
> RA_THRESHOLD
> DBSPACETEMP tempspace
> FILLFACTOR 90
> USEOSTIME 0
> MAX_PDQPRIORITY 100
> DS_MAX_QUERIES
> DS_TOTAL_MEMORY
> DS_MAX_SCANS 1048576
> DATASKIP off
> OPTCOMPIND 0
> ONDBSPACEDOWN 2> LBU_PRESERVE 0
> OPCACHEMAX 0
> HETERO_COMMIT 0
> OPT_GOAL -1
> DIRECTIVES 1
> RESTARTABLE_RESTORE off>
>
>
>