Timesten replication delay

Hi,
1. what is the utility to find the timesten latency, in otherwords how behind one timesten instance is behind the other one? in this case is there a configurable thresold that can be used to monitor this delay and generate alarm?
2. When we have timesten replication running between two timesten instances, due to some reason if the other instance of timsten is down for some time and once it comes up and replication is started, what is utility which can be used to find that both the instances are in sync.?
any procedure other then finding the count in the table:)

There are several ways to view/monitor the replication latency, or lag. ttRepAdmin -showstatus will display latency information. The builtin procedures 'ttBookmark' and ttLogHolds' will show information on replication bookmarks and hence replication backlog. You need to implement a monitoring process that uses these metrics to raise an alarm if the backlog becomes too large; TimesTen does not have any automatic way to do this currently.
Jeremy already answered question (2).
Chris

Similar Messages

Timesten replication with multiple interfaces sharing the same hostname

Hi,
we have in our environment two Sun T2000 nodes, running SunOS 5.10 and hosting a TT server currently in Release 7.0.5.9.0, replicated between each other.
I would like to have some more information on the behavior of the replication w.r.t. network reliability when using two interfaces associated to the same hostname, the one used to define the replication element.
To make an example we have our nodes sharing this common /etc/hosts elements:
151.98.227.5 TBMAS10df2 TBMAS10df2-10 TBMAS10df2-ttrep
151.98.226.5 TBMAS10df2 TBMAS10df2-01 TBMAS10df2-ttrep
151.98.227.4 TBMAS9df1 TBMAS9df1-10 TBMAS9df1-ttrep
151.98.226.4 TBMAS9df1 TBMAS9df1-01 TBMAS9df1-ttrep
with the following element defined for replication:
ALTER REPLICATION REPLSCHEME
ADD ELEMENT HDF_GNP_CDPN_1 TABLE HDF_GNP_CDPN
CHECK CONFLICTS BY ROW TIMESTAMP
COLUMN ConflictResTimeStamp
REPORT TO '/sn/sps/HDF620/datamodel/tt41dataConflict.rpt'
MASTER tt41data ON "TBMAS9df1-ttrep"
SUBSCRIBER tt41data ON "TBMAS10df2-ttrep"
RETURN RECEIPT BY REQUEST
ADD ELEMENT HDF_GNP_CDPN_2 TABLE HDF_GNP_CDPN
CHECK CONFLICTS BY ROW TIMESTAMP
COLUMN ConflictResTimeStamp
REPORT TO '/sn/sps/HDF620/datamodel/tt41dataConflict.rpt'
MASTER tt41data ON "TBMAS10df2-ttrep"
SUBSCRIBER tt41data ON "TBMAS9df1-ttrep"
RETURN RECEIPT BY REQUEST;
On this subject moving from 6.0.x to 7.0.x there has been some changes I would like to better understand.
6.0.x reported in the documentation for Unix systems:
If a host contains multiple network interfaces (with different IP addresses),
TimesTen replication tries to connect to the IP addresses in the same order as
returned by the gethostbyname call. It will try to connect using the first address;
if a connection cannot be established, it tries the remaining addresses in order
until a connection is established.
Now On Solaris I don't know how to let gethostbyname return more than one interface (the documention notes at this point:
If you have multiple network interface cards (NICs), be sure that “multi
on” is specified in the /etc/host.conf file. Otherwise, gethostbyname will not
return multiple addresses).
But I understand this could be valid for Linux based systems not for Solaris.
Now if I properly understand the above, how was the 6.0.x able to realize the first interface in the list (using the same -ttrep hostname) was down and use the other, if gethostbyname was reporting only a single entry ?
Once upgraded to 7.0.x we realized the ADD ROUTE option was added to teach TT how to use different interfaces associated to the same hostname. In our environment we did not include this clause, but still the replication was working fine regardless of which interface we were bringing down.
My both questions in the end lead to the same doubt on which is the algorithm used by TT to reach the replicated node w.r.t. entries in the /etc/hosts.
Looking at the nodes I can see that by default both routes are being used:
TBMAS10df2:/-# netstat -an|grep "151.98.227."
151.98.225.104.45312 151.98.227.4.14000 1049792 0 1049800 0 ESTABLISHED
151.98.227.5.14005 151.98.227.4.47307 1049792 0 1049800 0 ESTABLISHED
151.98.227.5.14005 151.98.227.4.48230 1049792 0 1049800 0 ESTABLISHED
151.98.227.5.46050 151.98.227.4.14005 1049792 0 1049800 0 ESTABLISHED
TBMAS10df2:/-# netstat -an|grep "151.98.226."
151.98.226.5.14000 151.98.226.4.47699 1049792 0 1049800 0 ESTABLISHED
151.98.226.5.14005 151.98.226.4.47308 1049792 0 1049800 0 ESTABLISHED
151.98.226.5.44949 151.98.226.4.14005 1049792 0 1049800 0 ESTABLISHED
Tried to trace with ttTraceMon but once I brought down one of the interfaces did not see any reaction on either node, if you have some info it would be really appreciated !
Cheers,
Mike

Hi Chris,
Thanks for the reply, I have few more queries on this.
1.Using the ROUTE CLAUSE we can use multiple IPs using priority level set, so that if highest priority level set in thr ROUTE clause for the IP is not active it will fall back to the next level priority 2 set IP. But cant we use ROUTE clause to use the multiple route IPs for replication simultaneously?
2. can we execute multiple schema for the same DSN and replication scheme but with different replication route IPs?
for example:
At present on my system, I have a replication scheme running for a specific DSN with stand alone Master-Subscriber mechanism, with a specific route IP through VLAN-xxx for replication.
Now I want to create and start another replication scheme for the same DSN and replication mechanism with a different VLAN-yyy route IP to be used for replication in parallel to the existing replication scheme. without making any changes to the pre-existing replication scheme.
for the above scenarios, will there be any specific changes respective to the different replication schema mechanism ie., Active Standby and Standalone Master Subscriber mechanism etc.,
If so what are the steps. like how we need to change the existing schema?
Thanks In advance.
Naveen

TT12039: Could not get port number of TimesTen replication agent on remote

i use the "ttRepAdmin -duplicate -from yymhcc_active -host mt2 -setMasterRepStart -uid musicclub -pwd musicclub -remoteDaemonPort 17001 -keepCG -cacheUid musicclub -cachePwd musicclub -localhost "mt4" yymhcc_standby;"
TT12039: Could not get port number of TimesTen replication agent on remote host. Either the replication agent was not started, or it was just started and has not communicated its port number to the TimesTen daemon
at the mt4 machine :
ping mt2
PING mt2 (10.25.71.26) 56(84) bytes of data.
64 bytes from mt2 (10.25.71.26): icmp_seq=1 ttl=64 time=0.138 ms
64 bytes from mt2 (10.25.71.26): icmp_seq=2 ttl=64 time=0.108 ms
at mt2
ttAdmin -query yymhcc_active
RAM Residence Policy : manual
Manually Loaded In RAM : True
Replication Agent Policy : manual
Replication Manually Started : True
Cache Agent Policy : manual
Cache Agent Manually Started : True
and
repschemes;
Replication Scheme Active Standby:
Master Store: YYMHCC on MT2
Master Store: YYMHCC_STD on MT4
Excluded Tables:
None
Excluded Cache Groups:
None
Excluded sequences:
None
Store: YYMHCC on MT2
Port: 21000
Log Fail Threshold: (none)
Retry Timeout: 30 seconds
Compress Traffic: Disabled
Store: YYMHCC_STD on MT4
Port: 20000
Log Fail Threshold: (none)
Retry Timeout: 30 seconds
Compress Traffic: Disabled
1 replication scheme found.

when i use: ttRepAdmin -duplicate -from yymhcc -host mt2 -setMasterRepStart -uid musicclub -pwd musicclub -remoteDaemonPort 17001 -keepCG -cacheUid musicclub -cachePwd musicclub -localhost "mt4" yymhcc_standby
TT8179: Cannot create duplicate store : store already exists
master odbc config :
[yymhcc_active]
Description=For Active Master
Driver=/usr/local/timesten/TimesTen/tt70/lib/libtten.so
DataStore=/usr/local/timesten/TimesTen/yymhcc/yymhcc
DatabaseCharacterSet=ZHS16GBK
ConnectionCharacterSet=ZHS16GBK
Authenticate=1
OracleID=yymhcc
OraclePWD=mc
UID=mc
PWD=mc
#ipcs memory size(M)
PermSize=8192
Connections=2047
#permsize*20%
TempSize=1024
PassThrough=1
WaitForConnect=0
Isolation=0
Logging=1
DurableCommits=0
CkptFrequency=600
CkptLogVolume=256
#LogBuffSize=256000
LogBuffSize=524288
LogFileSize=256
LogFlushMethod=1
LogPurge=1
LockLevel=0
LockWait=5
SQLQueryTimeout=5
RecoveryThreads=16
subscriber config :
[yymhcc_standby]
Driver=/usr/local/timesten/TimesTen/tt70/lib/libtten.so
#DataStore=/usr/local/timesten/TimesTen/yymhcc_std/yymhstd
DatabaseCharacterSet=ZHS16GBK
ConnectionCharacterSet=ZHS16GBK
Authenticate=1
OracleID=yymhcc
OraclePWD=mc
UID=mc
PWD=mc
PermSize=8192
Connections=2047
#permsize*20%
TempSize=1024
PassThrough=1
WaitForConnect=0
Isolation=0
Logging=1
DurableCommits=0
CkptFrequency=600
CkptLogVolume=256
#LogBuffSize=256000
LogBuffSize=524288
LogFileSize=256
LogFlushMethod=1

TimesTen Replication with Oracle

I am generating some 100k data in TimesTen i would like to replicate those data to Oracle database.
What will happen if there is sudden crash in the system of TimesTen ,how replication will takes place ,once again i reconfigure the replication to oracle whether the data which are inserted in TimesTen will get reflected or i need to manually need to load the data,is it is possible.
If there is sudden crash then how the data at the TimesTen end will be written to the disk ,since data which are been written to disk are on the basis of check-pointing and transaction logging. if so how to avoid this situation in real time.
Thanks!

The 'replication' to Oracle (actually AWT cache group) works by capturing the changes from the TimesTen transaction logs. Once a change reaches the TT transaction logs on disk it is safe and once the TT machine, TimesTen and the replication agent are back up replication will resume from where it left off.
More generally, data stored in TimesTen is persisted and protected using the checkpointing and transaction logging mechanism within TimesTen. By default, TimesTen uses asynchronous write to the logs (DurableCommits=0). This gives very high performance but in the event of a failure you will lose the last few committed transactions. If this is not acceptabel you can tell TimesTen to use synchronous commit to disk (DurableCommits=1). This is much (and I mean hugely) slower but ensures that all committed transactions are safe (unless the disk itself fails). TimesTen also provides TimesTen to TimesTen replication (both asynchronous and synchronous) as an additional / alternative method of data protection.
Chris

Replication delay is increasing for every 30sec ,how to reduce this

ReplicaDn Consumer Supplier Delay Transport
o=domain.com ldap01.domain.com:3877 server1.domain.com:3877 6158 SSL 07/02/09 00:05:09
= ldap02.domain.com:3877 server1.domain.com:3877 6158 SSL 07/02/09 00:05:09
= ldap03.domain.com:3877 server1.domain.com:3877 6158 SSL 07/02/09 00:05:09
After adding 2 HUB servers and 1 master to master server ,replication delay is increasing for every 30se when I ran insyn command. I tried all the options send updates /forcing the things but still no luck ,please let me know what should I do to reduce this.to make 0

we are using 5.2 P4 . After involving SUN engineer ,he said changelogdb has been corrupted and needs to be re-initialized all the other 5 servers from this source ldap server. We did that ,after that all servers are in SYNC now.
Just wondering when doing promotion/demotion of the servers out of 6 (4mmr + 2 hubs) do we have to follow any precision specific steps ?
I followed these simple steps. 1. demoted 2 masters out of 4 masters to hubs
2. deleted all references to these hubs from other servers
3. Promoted the old existing hubs to masters and created the repl agreements
while doing this exercise found out changelogdb corrupted ,resulting all servers had to be re-initialized (which is the dreadiest thing I wanted to avoid) .
Any points am missing from this promotion/demotion process ?

Timesten Replication: Subscriber DB is necessary?

Hi all,
I learned about Timesten replication, but I didn't understand clearly the purpose of Subscriber DB in Replication architecture.
I think pair active and standby is enough for durability and availability of data.
Subscriber DB is necessary?
Thanks,
LeoBon

Subscriberr DBs can also be used for 'read scaling'. If the application has modest update rates but needs a very high level of read throughput distributed acrosss multiple machiens then the A/S pair can be used as a high availability provisioning point . Read access can be handled by both the A/S pair and one or more subscribers. This is a very easy way to scale out to support a heavy read workload.
Subscriber DBs are optional not mandatory.
Chris

Windows 2008 FRS port Customization causing Replication Delay

Hi
we have 4 windows 2008 R2 DC. we have restricted the port for FRS Replication. now we are facing replication delay.
Not finding any error in DCDIAG, REPADMIN /Showreps and in FRS event there is an error 13508 and it is not recovering since from 3 days to 13509 so please suggest the solution
whether Port customizion is recommended

Hi,
We can restrict FRS replication traffic to a specific static port. However, if there is an intermediate network device or software that is being used to filter packets between domain controllers, we need to verify that the device or the software
allows communication over the specified port. Additionally, make sure the destination TCP port that we set is open on the firewall.
Regarding restricting FRS replication traffic to a specific static port, the following article can be referred to for more information.
How to restrict FRS replication traffic to a specific static port
https://support.microsoft.com/en-us/kb/319553
Regarding troubleshooting FRS Event ID 13508, the following article can be referred to for more information.
Troubleshooting File Replication Service
https://msdn.microsoft.com/en-us/library/bb727056.aspx#EMAA
Best regards,
Frank Shen
Please remember to mark the replies as answers if they help and unmark them if they provide no help. If you have feedback for TechNet Subscriber Support, contact [email protected]

Monitoring the health of TimesTen replication

Hello,
I have question on following,
Execute the log analyze command against a particular data store to generate the following information:
The number of transactions that are waiting to be replicated. For each transaction that has not been replicated, the information collected includes the number of operations for each transaction and the total size of each transaction (including partial rollbacks).
when i execute help on ttXactLog it doesn't show me the option of -logAnalyze, what is alternative for the number of transactions/logs waiting to be replicated?
ttXactLog -help
** This utility is primarily intended for use by TimesTen customer **
** support to diagnose problems. This information may also be      **
** useful to system administrators and developers.                 **
Usage:
ttXactLog {-h | -help | -?}
ttXactLog {-V | -version}
ttXactLog [-v <verbosity>] [-m <maxChars>] [-s] [-t] [-b <blkId>]
            [-l1 <lfn>.<lfo> [-l2 <lfn>.<lfo>]] [-r <recType>] [...]
            [-tr <dir>] [-lb] [-headers <recs>] [-logdir <dir>]
            {<DSN> | -connstr <connStr> | <dsname>}

Sorry for the delay in replying.
What version of TimesTen are you running? the -logAnalyze option is only available in recent versions.
In older releases you can get some information on the replication backlog using
ttRepAdmin -bookmark DSN
but this only gives an indication of how much transaction log is queued; it is not as precise as the -logAnalyze function. If you need -logAnalyze then you need to upgrade to a recent Timesten version.
Chris

Is Multimaster Timesten replication a good option for huge data volumes?

Hi,
There are 3 timesten nodes in our production setup .There will be around 5 million rows in each node initially which will gradually increase to about 10 million.Once our application moves to production, there will be around 50-70 transactions per second in each node , which need to be replicated on to the other node.
Initially we thought of going with Active-Standby-Subscriber replication.However in this case if active and standby node go down,then it will be a site failure case.So is Active-Active (Multimaster replication) configuration a good option ? Will data collision happen when replication happens in both directions?
Thanks in advance.
Nithya

Multi-master replication is rarely a good idea. You will get data collisions unless you rigorously partition the workload. Conflict detection and resolution is not adequate to guarantee consistency over time. Recovery back to a consistent state after a failure is complex and error prone. I'd strongly advise against a multi-master setup, especially for a high volume system.
You seem to be concerned that 2 out of the 3 systems may fail resulting in a site outage. The likelihood of that is small if you have set things up with separate power etc. With the A/S pair based approach you would still have query capability if the two master systems failed. The chances of all 3 systems failing is not that much less than of just 2 failing in reality I would say (depending on the reason for the failure).
Chris

TimesTen Replication Question

We have a customer that is experiencing some odd behavior with their TimesTen 7.0.3.0 install. Once in a while (every 3 or 4 days) the replication state (as shown by the ttRepAdmin command) goes into PAUSE state momentarily and at the same time the machine experiences a CPU spike. As far as I know the only way the replication state will change is via the ttRepAdmin command, meaning it would have to be done manually or by some process calling the ttRepAdmin command. They are certain that this is not being done manually and I’m certain that none of our processes is doing it.
Does anyone know if there is any way that replication state goes from START to PAUSE and back to START on its own? As for the CPU spikes, my guess is that this is caused by the state transition from PAUSE to START.
Thanks.

There is nothing within TimesTen that will set the replication state to PAUSE but it could be a bug in the reporting of the state. Do these events coincide with any other events happenning on the system? Are there any messages reported in the TimesTen daemon log when this occurs? Have you observed what process(es) use more CPU when the CPU spike occurs? We've never heard of anything like this before so investigation will be needed to determine what is going on.
Chris

TimesTen Replication: Switching users on Failover.

Hello, I have installed TimesTen on two machin and created bidirectional replication scheme for setup master/hotstandby connection.
Not very sure how TimesTen standby setting should be done.
My understanding was once I create the master/hotstandby, the hotstandby will not accept the connections and need to activate it once master goes down.
But when I setup the replication scheme I found that users can connect to master as well as hotstandby server.
Is it the way hotstandby work? I mean master and standby both are active?
If the master crashes how the users will be transferred to standby.
Do we need to activate the standby server when master is crashed?

Hello Chris, Thanks for your feedback. I have created active standby pair as follows:
1. Created DSN (ActiveMaster) on Master and (ActiveStandby) on Standby Server.
2. Created USER ttactive on Master and Standby Server and Oracle.
3. Created tables in Oracle10g.
4. Created Cachegroup in Master.
5. Created active Standby on master as follows:
CREATE ACTIVE STANDBY PAIR
ActiveMaster ON "hostnames masterserver",
ActiveStandby ON "hostname standbyserver"
RETURN RECEIPT
STORE ActiveMaster PORT 20000 TIMEOUT 120
STORE ActiveStandby PORT 21000 TIMEOUT 120;
6. Started replication agent on Master. (ttAdmin –RepStart ActiveMaster)
7. Trying to create duplicate datastore for standby using ttRepAdmin -duplicate.
>ttRepAdmin -dsn ActiveStandby -duplicate -from ActiveMaster -
host "hostname masterserver" –UID ttactive –PWD tt123
-setMasterRepStart - keepCG
====>> This utility is failing with error. "TT12078: Failed to reset
is_local_store TT12078: TT15001: User lacks privilege WRITE -- file "comp.c",
lineno 4369, procedure "sbPtCheckPriv". File: repSelf.c, line: 899"
8. Since the -duplicate did not work. I created the AWT cachegroup in standby same as master and created the replication scheme same as master.
9. I executed ttRepStateSet('ACTIVE') on ActiveMaster DataSource.
Issues:
1. On Checking the status on Master and Active using ttRepStateGet.
Master shows "ACTIVE" but StandbyShows "IDLE"
2. When I insert any record in Master it comes back with no receipt error.
3. Looking at cachgroup tables I observe following line after each table in cg.
15001: User lacks privilege SELECT
Table Type: Propagate
What does this mean? Where I need to give the privileges.
Could you please guide me how to go ahead with it.
Best Regards
Tanweer Ahmed Khan.

Timesten to Timesten replication error

Hi,
I am testing timesten db. I installed the db on two windows 2000 servers and I set the DSN client and Data manager settings. The server can connect each other db via DSN client. I tried to set up replication timesten to timesten but I get error. The error below and any advice would be helpful
Thanks
Server1
DSN Name TODB (Timesten Database Manager)
DB Location c:\OnlineDB
Log Location c:\OnlineDB\log
Database Character set =UTF8
DB Permanent size 20
DB Temp Permanent size 20
Connection Character Set=UTF8
Server2
DSN Name FODB (Database Manager)
DB Location c:\OnlineDB
Log Location c:\OnlineDB\log
Database Character set =UTF8
DB Permanent size 20
DB Temp Permanent size 20
Connection Character Set=UTF8
Server1
Client DSN FODB (Timesten Client)
ServerName Server2
Server DSN FODB
Connection Character Set=UTF8
user:test
pass:test
( I could connect timesten client after create an user and grant all to that user on server2)
a sql code run into two server (TODB and FODB)
connect TODB;
CREATE TABLE repl.tab (a NUMBER NOT NULL,
b NUMBER,
c CHAR(8),
PRIMARY KEY (a));
connect FODB;
CREATE TABLE repl.tab (a NUMBER NOT NULL,
b NUMBER,
c CHAR(8),
PRIMARY KEY (a));
connect TODB;
CREATE REPLICATION repscheme
ELEMENT e TABLE repl.tab
MASTER TODB
SUBSCRIBER FODB;
I get the error after the create replicaton sql codes run.
"SQL Error: [TimesTen][TimesTen 7.0.3.0.0 ODBC Driver][TimesTen]TT17029: This store (ONLINEDB on Server1) is not involved as a local store in replication scheme ADMINISTRATOR.REPSCHEME -- file "eeRepDDL.c", lineno 3214, procedure "sbEeRepCreateEval()""

You may get your answer faster if you post this question to TimesTen Discussion Forum:
TimesTen In-Memory Database

TimesTen replication port setup

Hi,
I already had replication Timesten to Timesten working using the default port. (TimesTen Version: 7.0.5.0.0).
Now that I specified one port (16300) to be used, I receive an error: code 12039 when trying to duplicate using the C API.
==========
ttwrapper::ttRepDuplicateEx
==========
The only thing that I've changed was to include "PORT 16300" in the replication script below:
=============================================
autocommit 0;
{CALL ttRepPolicySet( 'manual' )};
{CALL ttRepStop};
DROP REPLICATION xxx;
commit;
{CALL ttCompact};
{CALL ttCkpt};
{CALL ttCkpt};
commit;
CREATE REPLICATION xxxl
ELEMENT A DATASTORE
MASTER "dsn_name" ON "master_hostname"
SUBSCRIBER "dsn_name" ON "subscriber_hostname"
ELEMENT B DATASTORE
MASTER "dsn_name" ON "subscriber_hostname"
SUBSCRIBER "dsn_name" ON "master_hostname"
STORE "dsn_name" ON "master_hostname" PORT 16300 FAILTHRESHOLD 200
STORE "dsn_name" ON "subscriber_hostname" PORT 16300 FAILTHRESHOLD 200;
commit;
=============================================
What else do I need to setup?
Thanks,
Milena

Aha; this is your problem.
You cannot use the same port number for more than one DSN located in the same machine, even if they are in different TimesTen instances and/or are different TimesTen versions. The combination of hostname(IP address) and port number must be unique within a machine and if you try and use the same port number for multiple replication schemes then multiple processes will be trying to bind to the same well known port. I suspect that if you look you will see errors in the main daemon log from one of the replication agents.
The port number you use must be unique within the machine; not only must it not be used by any other TimesTen process 9replication, server, main daemon) but it must not be in use by any non-TimesTen process either.
Chris

Replication delay

Lets say you have a writer and a number of replicas in a BerkeleyDB system.
When a writer performs an update, are this update synchronized onto all replicas and are performed there,
or does BerkeleyDB put that in some sort of background process so that it can take care of the next update
to increase response time ?
I'm asking just be sure I got this right.
//

It isn't exactly clear what you are asking.
There is no sort of background process sending the messages. If you are using the
RepMgr API the thread sending the message sends it right away. If you are using
the Base API, the application is, of course, free to do whatever it wants. In that case,
each message that needs to be sent without any delay is marked as DB_REP_NOBUFFER.
When updates occur, those updates are only applied on the client replica sites at the
time that the txn commit happens.
The synchronization aspect of updates is up to the application. Any site can choose to
set DB_TXN_NOSYNC or not. If you are using the RepMgr API, you can select an
acknowledgement policy that suits the needs of your application. The master, then, will
or will not wait for some number of clients to acknowledge the successful application
of a txn based on that policy.
Sue LoVerso
Oracle

Replication port used in TimesTen 5.1.35

Hi,
Can you please tell me what is the default port number used by TimesTen replication in version 5.1.35 ?
Regards
Pratheej

If you're using AUTO port assignment for replication (ie not specifying a PORT in your Create Replication SQL), then use ttRepAdmin -showstatus to see what the port is that's been assigned by the OS. This is also specified in the timestend.log when the replication threads are starting up. There isn't a default port used otherwise. The agent will just use whatever it's assigned by the OS on each startup.

Timesten replication delay

Similar Messages

Maybe you are looking for