Hi,

i am working with ER replication.
I have 2 computers :
A 4 CPU RM400 (siemens) with sinix OS (or reliant unix) and a 7.31UD2
database server.
A 8 CPU SunFire 880 with Solaris 8 and a 9.30UC2 database server.
The network is a 100Mbit lan.

All the database user tables of RM400 are replicated on the SunFire. So we
are working with master / slave configuration

But we have performance problems.
We are working only with queuemem and we stop replication when this one is
over than 45% filled.

On RM400 we use 3 CPU for informix and on SunFire we can use from 1 to 7
CPU.

On RM400 evaluator thread are faster than Sender thread so the Queuemem is
filling.
There is only one sender thread, can i add one and how to proceed ? How to
speed up this thread ?
Is it normal that if queuemem is filling, CPU used increases for the same
transaction load ?

When i do an onstat -g ntu command, i have the quota q-exceed that
increases. I modified SQLHOSTS file with b=16384 option on network
connections but it doesn't modify anything.
What is the real meaning of this quota and what is the impact on the sender
?

onstat -c |grep CDR on RM400 :
[informix:tests3_shm SDP21] $ onstat -c |grep CDR
# CDR Variables
CDR_LOGBUFFERS 2048 # size of log reading buffer pool (Kbytes)
#CDR_LOGBUFFERS 20480 # size of log reading buffer pool (Kbytes)
#CDR_EVALTHREADS 1,2 # evaluator threads (per-cpu-vp,additional)
#CDR_EVALTHREADS 4 # evaluator threads (per-cpu-vp,additional)
CDR_EVALTHREADS 0,1 # evaluator threads
(per-cpu-vp,additional)
CDR_DSLOCKWAIT 5 # DS lockwait timeout (seconds)
#CDR_QUEUEMEM 4096 # Maximum amount of memory for any CDR
queue (Kbytes)
#CDR_QUEUEMEM 32768 # Maximum amount of memory for any CDR
queue (Kbytes)
CDR_QUEUEMEM 131072 # Maximum amount of memory for any CDR
queue (Kbytes)
CDR_LOGDELTA 40 # % of log space allowed in queue memory
#CDR_NUMCONNECT 16 # Expected connections per server
CDR_NUMCONNECT 3 # Expected connections per server
CDR_NIFRETRY 300 # Connection retry (seconds)
CDR_NIFCOMPRESS 0 # Link level compression (-1 never, 0 none,
9 max)


onstat -c |grep CDR on SunFire
drc11:[informix]>onstat -c |grep CDR
# CDR Variables
#CDR_EVALTHREADS 1,2 # evaluator threads (per-cpu-vp,additional)
CDR_EVALTHREADS 1,0 # evaluator threads (per-cpu-vp,additional)
CDR_DSLOCKWAIT 5 # DS lockwait timeout (seconds)
#CDR_QUEUEMEM 4096 # Maximum amount of memory for any CDR queue
(Kbytes)
CDR_QUEUEMEM 32768 # Maximum amount of memory for any CDR queue
(Kbytes)
CDR_NIFCOMPRESS 0 # Link level compression (-1 never, 0 none, 9 max)
CDR_SERIAL 0 # Serial Column Sequence
CDR_DBSPACE # dbspace for syscdr database
CDR_QHDR_DBSPACE # CDR queue dbspace (default same as catalog)
CDR_QDATA_SBSPACE receivequeue # CDR queue smart blob space
CDR_QDATA_SBFLAGS 0 # Log/no-log (default no log)

If more information needed, ask me.

If possible it will be better for me to speak French.
But by email, no problem with english.

Thanks

sending to informix-list