Topic
No replies
JJ@IBM
JJ@IBM
18 Posts
ACCEPTED ANSWER

Pinned topic Connection Manager Linux 11.70.FC4 / CSDK 3.70.FC4 ... Not quite right?

‏2012-02-21T17:24:27Z |
Hi,

So, fairly straight forward configuration, and yet now the "usage" has changed to even more un-decipherable, and the manual has references in some places to "pre 11.70.xC3" and then in others to "post 11.70.xC3".

So, what is "wrong" with the following setup (if anything) ... basing things on the following sqlhosts file:
............................
jj_1170f_rep_1_pri_s onipcshm jjlinuxrepdemo jj_1170f_rep_1_pri_s
jj_1170f_rep_1_sec_s onipcshm jjlinuxrepdemo jj_1170f_rep_1_sec_s
jj_1170f_rep_1_rss1_s onipcshm jjlinuxrepdemo jj_1170f_rep_1_rss1_s
jj_1170f_rep_1_rss2_s onipcshm jjlinuxrepdemo jj_1170f_rep_1_rss2_s
jj_1170f_rep_1_sds1_s onipcshm jjlinuxrepdemo jj_1170f_rep_1_sds1_s
jj_1170f_rep_1_sds2_s onipcshm jjlinuxrepdemo jj_1170f_rep_1_sds2_s
jj_1170f_rep_1 group - - i=1
jj_1170f_rep_1_pri_t onsoctcp jjlinuxrepdemo jj_1170f_rep_1_pri_t g=jj_1170f_rep_1
jj_1170f_rep_1_sec_t onsoctcp jjlinuxrepdemo jj_1170f_rep_1_sec_t g=jj_1170f_rep_1
jj_1170f_rep_1_rss1_t onsoctcp jjlinuxrepdemo jj_1170f_rep_1_rss1_t g=jj_1170f_rep_1
jj_1170f_rep_1_rss2_t onsoctcp jjlinuxrepdemo jj_1170f_rep_1_rss2_t g=jj_1170f_rep_1
jj_1170f_rep_1_sds1_t onsoctcp jjlinuxrepdemo jj_1170f_rep_1_sds1_t g=jj_1170f_rep_1
jj_1170f_rep_1_sds2_t onsoctcp jjlinuxrepdemo jj_1170f_rep_1_sds2_t g=jj_1170f_rep_1
jj_1170f_rep_2_pri_s onipcshm jjlinuxrepdemo jj_1170f_rep_2_pri_s
jj_1170f_rep_2_sec_s onipcshm jjlinuxrepdemo jj_1170f_rep_2_sec_s
jj_1170f_rep_2_rss1_s onipcshm jjlinuxrepdemo jj_1170f_rep_2_rss1_s
jj_1170f_rep_2_rss2_s onipcshm jjlinuxrepdemo jj_1170f_rep_2_rss2_s
jj_1170f_rep_2_sds1_s onipcshm jjlinuxrepdemo jj_1170f_rep_2_sds1_s
jj_1170f_rep_2_sds2_s onipcshm jjlinuxrepdemo jj_1170f_rep_2_sds2_s
jj_1170f_rep_2 group - - i=2
jj_1170f_rep_2_pri_t onsoctcp jjlinuxrepdemo jj_1170f_rep_2_pri_t g=jj_1170f_rep_2
jj_1170f_rep_2_sec_t onsoctcp jjlinuxrepdemo jj_1170f_rep_2_sec_t g=jj_1170f_rep_2
jj_1170f_rep_2_rss1_t onsoctcp jjlinuxrepdemo jj_1170f_rep_2_rss1_t g=jj_1170f_rep_2
jj_1170f_rep_2_rss2_t onsoctcp jjlinuxrepdemo jj_1170f_rep_2_rss2_t g=jj_1170f_rep_2
jj_1170f_rep_2_sds1_t onsoctcp jjlinuxrepdemo jj_1170f_rep_2_sds1_t g=jj_1170f_rep_2
jj_1170f_rep_2_sds2_t onsoctcp jjlinuxrepdemo jj_1170f_rep_2_sds2_t g=jj_1170f_rep_2
  1. Connection manager specific entries
jj_1170f_rep1_oltp onsoctcp jjlinuxrepdemo jj_1170f_rep1_oltp
jj_1170f_rep1_mis onsoctcp jjlinuxrepdemo jj_1170f_rep1_mis
jj_1170f_rep1_sds onsoctcp jjlinuxrepdemo jj_1170f_rep1_sds
jj_1170f_rep2_oltp onsoctcp jjlinuxrepdemo jj_1170f_rep2_oltp
jj_1170f_rep2_mis onsoctcp jjlinuxrepdemo jj_1170f_rep2_mis
jj_1170f_rep2_sds onsoctcp jjlinuxrepdemo jj_1170f_rep2_sds
jj_1170f_replset_any onsoctcp jjlinuxrepdemo jj_1170f_replset_any
jj_1170f_grid_all onsoctcp jjlinuxrepdemo jj_1170f_grid_all

and the following CM configuration file:

NAME rep_cmconfig_1
MACRO jj_rep_1_servers=(jj_1170f_rep_1_pri_t,jj_1170f_rep_1_sec_t,jj_1170f_rep_1_rss1_t, \
jj_1170f_rep_1_rss2_t,jj_1170f_rep_1_sds1_t,jj_1170f_rep_1_sds2_t)
MACRO jj_rep_2_servers=(jj_1170f_rep_2_pri_t,jj_1170f_rep_2_sec_t,jj_1170f_rep_2_rss1_t, \
jj_1170f_rep_2_rss2_t,jj_1170f_rep_2_sds1_t,jj_1170f_rep_2_sds2_t)
LOG 1
DEBUG 0
LOGFILE /data/IBM/informix/spaces/jj_1170f_rep/CM/rep_cmconfig_1.log

CLUSTER jj_1170f_rep_1 # cluster name (unique in CM)
{
INFORMIXSERVER ${jj_rep_1_servers}
SLA jj_1170f_rep1_oltp DBSERVERS=primary
SLA jj_1170f_rep1_mis DBSERVERS=(HDR,RSS)
SLA jj_1170f_rep1_sds DBSERVERS=SDS
FOC ORDER=SDS,HDR,RSS TIMEOUT=5 RETRY=1
# CMALARMPROGRAM ${INFORMIXDIR}/etc/CMALARMPROGRAM.sh
}
CLUSTER jj_1170f_rep_2 # cluster name in CM
{
INFORMIXSERVER ${jj_rep_2_servers}
SLA jj_1170f_rep2_oltp DBSERVERS=primary
SLA jj_1170f_rep2_mis DBSERVERS=(HDR,RSS)
SLA jj_1170f_rep2_sds DBSERVERS=SDS
FOC ORDER=,SDS,HDR,RSS TIMEOUT=5 RETRY=1
# CMALARMPROGRAM ${INFORMIXDIR}/etc/CMALARMPROGRAM.sh
}
  1. Unable to get REPLSET CM working
REPLSET jj_1170f_replset # replset name
{
INFORMIXSERVER jj_1170f_rep_1_pri_t,jj_1170f_rep_2_pri_t
SLA jj_1170f_replset_any DBSERVERS=ANY # node discovery
}
  1. Unable to get GRID CM working
GRID jj_1170f_grid # grid name
{
INFORMIXSERVER jj_1170f_rep_1,jj_1170f_rep_2
SLA jj_1170f_grid_any DBSERVERS=ANY POLICY=LATENCY
}
............................

> cdr list grid jj_1170f_grid
Grid Node User
------------------------
jj_1170f_grid jj_1170f_rep_1* informix
jj_1170f_rep_2

Connection Manager started with INFORMIXSERVER set to jj_1170f_rep_1_pri_t

1. Connection Manager takes 12 seconds to register.

2. Both cluster entries work, but whenever shutting down CM, entry in online log ...
16:53:16 CM:Connection manager rep_cmconfig_1 registered with the server
16:54:26 CM:Session for Connection manager rep_cmconfig_1 terminated abnormally

3. Both cluster entries work ... for a while ... and then CM deregisters abnormally ... and then registers again!
17:13:09 CM:Connection manager rep_cmconfig_1 registered with the server
17:14:21 CM:Session for Connection manager rep_cmconfig_1 terminated abnormally
17:14:53 Checkpoint Completed: duration was 1 seconds.
17:14:53 Tue Feb 21 - loguniq 7, logpos 0x613018, timestamp: 0x1b6d45 Interval: 136
17:14:53 Maximum server connections 6
17:14:53 Checkpoint Statistics - Avg. Txn Block Time 0.000, # Txns blocked 0, Plog used 11, Llog used 12
17:15:57 CM:Connection manager rep_cmconfig_1 registered with the server

It appears to get confused when there are two different clusters - sometimes connecting to both (correctly) and then sometimes failing on one and then connecting to the other etc. etc.

4. Connecting to REPLSET SLA jj_1170f_replset fails with 908

5. Connecting to GRID SLA jj_1170f_grid_all fails with 908

6. The examples are pretty "short" and not complete examples.

JJ

"You are what you is"