how many transactions per second transactional replication can handle?
Server has 128GB RAM, 12 core 2 Process (24 core total) each transaction is approximately 32k.
Publisher and Subscriber are on different servers on same LAN, separate disks for logs, data, indexes and tempDB.
I've found data on how many transactions per second SQL Server can handle but nothing on replication.
On SQL 7.0, NT SP6 we had transactional replication running fine last Friday. After a reboot (Fri night) replication continued fine and then by Tuesday am, we saw a lot of 'timeout errors' for the distributor agent and the difference in numbers of rows (count(*)) for same tables on each side grew alarmingly high even though the replication continued fine..
I have changed the timeout from '0' to '300' which seemed to help with timeouts, but the difference in row counts kept progressing slowly while the replication seemed to be running fine..
Eventually we resorted to re-intializing the subscriptions on the replicated database to get past the issue and now 24 hours later after intializing was complete, we are still pretty much in synch..
Any one have any ideas why we got so much out of synch (wrt row counts) .. ? Any bugs/ gotchas/ quirks any one can think of?
thanks!!
Bruce
Hi,
i have a publisher, a distributer and 3 subscribers all working. All created at the same time.
Today i have tried to introduce a 4 th subscriber server and replication has fallen over. Each cycle will have updates eg
merged 75 data changes(6 inserts, 6 updates, 63 deletes, 48 resolved conflicts)
but the next cycle will show:
merged 55 data changes(4 inserts, 3 updates, 48 deletes, 37 resolved conflicts)
the 48 deletes are the conflicts from the previous cycle. this repeats hour on hour. The conflicts table is storing all of the changes but what have i done wrong in introducing the 4th server to get these symptoms? I followed the wizard all the way (may be that was it? :rolleyes: ) i have disabled this agent now and the others are working fine again. it is merge replication although the subscriber i have added is not going to be causing any updates.
conflicts are set to be resolved automatically
I have configured a transactional replication between 2 MSSQL 2005 instances. It is set to replicate every 15 minutes. Most replications take less than a minute with about 10 - 50 transactions being replicated.
However, 3 times during the day, the replication takes about 25 minutes to complete with 500-700 transactions being replicated. This also causes the application that is accessing the DB being replicated to become unresponsive when it attempts to retrieve rows from tables in the DB.
Might there be a setting that causes the transactional replication to accumulate transactions during the day?
Hi all,Does anyone know where I can find an implementation of the TPC-Cbenchmark for SQL Server 2000 (preferably written in .NET) ? All I canfind is an old Microsoft implementation of TPC-B written in C, and myC skills are rusty to say the least :( Oracle seems to have Hammerora,so maybe someone ported it to SQL Server ? In the worst case, ifanyone would be interested in collaborating on writing such animplementation, I'd be more than happy to get it started.Thanks,Oliver
View 6 Replies View RelatedI am beginning design for a project to maintain an on-line archive of historical data for a production database, and then truncate old, rarely-used data from the production database. I am considering a method whereby we (A) replicate data changes from the production database to an archive database and (B) truncate old data from the production database according to predetermined business rules.
I am trying to determine whether SQL Server merge or transactional replication can be used to accomplish the replication portion of this requirement.
Given this scenario, there are two types of modifications that are made to the production database: (1) application inserts, deletes and updates which should be replicated, and (2) truncation deletes (as described in "B", above) which should not be replicated.
My question is: Will I be able to control the replication process such that application deletes are replicated, but "truncation" deletes are not? Could static row filters somehow be used to accomplish this?
Hi Guys,
We have developed some applications and we wanted to have a performance bench mark for the database. Can anyone pl let me know how to have a baseline for performance. What would be and acceptable response time. Thanks for your reply.
--Joy
Hi,
Has anyone seen any benchmark results for SSIS vs Informatica?
Thanks,
David
Mariner
I'm trying to create a transaction replication from SQL Server 2000 to 2005. Basic replication between the servers works just fine. However, what I want to accomplish is to be able to skip some of the transactions. Example - from time to time we want to purge some of the historical data from the main database (the publisher). We don't want the same purging to occur on the destination database, which will be used for reporting purposes and needs to include all the historical information. I wanted to simply stop the replication log reader, purge the records, backup the transaction log with truncation and then restart the reader. The only problem - the truncation on the replicated database keeps the transactions of the purging until they are replicated, so the transaction log backup doesn't help. Any ideas would be greatly appreciated!
View 5 Replies View RelatedHow cheap (best free) database benchmark do you recommend me for SQL Server 2005(Express Edition)?
View 2 Replies View RelatedIn an instance with 2008R2, the value of Transactions/sec in Master DB is very bigger than an instance with 2012.
Is it normal? I ask it because both instances execute the same application with the normals differences between them (the Transactions/sec's number for the application's bds is similar).
There is a database "Foo" sitting on server "A". There is a database "Bar" sitting on server "B". A.Foo publishes a subset of its schema. B.Bar subscribes to A.Foo's publication. The distribution database is on "B" (B.distributor). This a push subscription (transactions are pushed to the subscriber from the distributor). Every day (including the weekend) I get the following alert:
"5/12/2015 3:53:16 AM, Unsubscribed Transactions (Count) on "B" is Warning.
SQL Server instance "B" has 636771 unsubscribed replication transactions received by the Distributor and not received by a Subscriber.
Unsubscribed Transactions (Count): Number of replication transactions received by the Distributor and not received by a Subscriber."
The number of transactions will vary. The alerts will be sent between 1:20 AM (EST) and 3:30 AM (EST). I'm trying to figure out what is causing the backup of transactions. I assume the issue precedes the alerts by 30-minutes or so.
There are no backups occurringNothing is blocking the distributor agent in the subscription databaseJob activity is at a minimum; the few jobs running run throughout the dayThe machine has plenty of resources -- CPU, RAM, etc.The publisher database shows no signs of stress.
Recently we disabled the transactional replication on a prodution database(data size:300GB) , two weeks later we were aware of the transaction log of that db almost reached the limitation (70GB).I investigated the issues and found:
1. open transactions pending for distribution:
DBCC opentran('xxxx')
Replicated Transaction Information:
Oldest distributed LSN : (0:0:0)
Oldest non-distributed LSN : (323830:7083:1)
DBCC execution completed. If DBCC printed error messages, contact your system administrator.
2. exec sp_replcounters, got over 240,000 replicated_transactions, which caused the tran log extremely large.
The transactional replication has been disabled on this server, there are no publication, distribution DBs any longer, all replication-related agents are also gone, why did this still happen???
3. use sp_repldone manaully reset those 240,000 transactions to distributed, check opentran, it was fine for a while. The tran log file was freed up 99% after that.
4. check open transactions and run sp_replcounters today, found the over 60000 transactions again on the db.
Did anybody experience the same issue and knows how to fix this issue and stop this happening ?
Many Thanks,
Henry
Hi,
I have multiple questions and would appreciate any suggestions in resolving them. I'm a novice to these issues.
1) First of all, what is the exact command to setup the trunc. log on chkpt. option on for a transactional log of a SQL Server 2000 database? Is this option on by default? I have noticed for one of the databases I'm managing that the transaction log was over 12 GB, while the db was only 425 MB.
2) How's it possible to run a DBCC TRACEON to see the content of the transaction log to see if we are having any issues with any uncommitted transactions, i.e. updates, inserts, and/or deletes.
3) What are the commands to truncate inactive transactions and increasing the readbatchsize?
4) lastly, how do I validate transactional replication via checksum and find valid latency between a small number of changes that need to be committed between publisher and subscriber.
Thank you so much in advance!
Sincerely,
Alla
I have a update trigger. In this trigger I need to insert few records in 3 tables. If error comes in any of these inserts then previous inserts to get committed. This trigger was written in Sybase and it was possible to create transaction and commit the transactions.
View 4 Replies View RelatedSo I've been monitoring long-running transactions on a SQL Server that hosts a couple of vendor-supplied databases that look after our factory.Today I noticed a pair that have confused my Excel spreadsheet (that I've been using to analyze these transactions).So here's the weird thing that I spotted. Given this query:
SELECT p.spid, p.login_time, at.transaction_begin_time, datediff(second, p.login_time, at.transaction_begin_time) as [difference]
FROM sys.sysprocesses AS p INNER JOIN
sys.dm_tran_session_transactions AS st ON st.session_id = p.spid INNER JOIN
sys.dm_tran_active_transactions AS at ON st.transaction_id = at.transaction_id
[code]....
I had a look in the event log on the server, which had just been rebooted at around that time. It seems that the clock got changed on boot-up, with the size of it quite surprising. This meant that these processes were able to start their transactions *before* they logged on. Hopefully this doesn't cause any other weird problems.So I've requested an investigation about time synchronization on our virtualization hosts... and in the mean time, have set the SQL Server services to 'delayed start'.
We have a filtered publisher (SQL Server 2012) with several pull and push subscriptions that have been functioning fine for years.
We have just added a new pull subscriber from a remote server. We configured an alternate folder location on the subscriber (pointing to existing path on a remote server). We generated new snapshot and partitioned snapshot, and moved to subscribing server at alternate path.We're seeing a problem we've never seen before:
When the merge agent runs, it applies the schema and performs a few hundred bulk inserts, but then proceeds to 'download changes to subscriber' as though the snapshot is old (it is brand new and should have all transactions). We have never seen this before, i.e. all data is usually applied in bulk when creating a new or reinitializing an existing subscriber.
we marked the subscriber for re-initialization (from both the publishing and subscribing server) with a fresh set of snapshots. The same problem occurred.We have also tried in our test environment (which mirrors production) and the snapshot applies as expected (via bulk inserts).
(It might be relevant to note the publisher was recently upgraded in-place from 2005 to 2012.)
I vaguely remember reading somewhere that all distributed transactions are executed at Serializable Isolation Level "under the covers."
1. Is this true?
2. What does "under the covers" mean in this case; i.e. will I not see the isolation level represented accurately in requests?
It is my understanding that when having LinkedServers, the option "enable promotion of Distributed Transactions for RPC" should be set to TRUE, so we can rollback , if needed, remote transactions. At least, that's my understanding of that setting.
Having said that, the TRUE setting is affecting this particular TSQL code, inside an sproc, which I would prefer not to alter:
Insert into #TempTable
EXEC ServerB.MyDatabase.MyStoreProcedure
@param1= '',
@param2= ''
When set is set to TRUE (current setting) I get this error:
OLE DB provider "SQLNCLI11" for linked server "ServerB" returned message "The partner transaction manager has disabled its support for remote/network transactions.".
Msg 7391, Level 16, State 2, Line 28
The operation could not be performed because OLE DB provider "SQLNCLI11" for linked server "ServerB" was unable to begin a distributed transaction.
... when set to off, the error goes away.
Hi there,
I have decided to move all my transaction handling from asp.net to stored procedures in a SQL Server 2000 database. I know the database is capable of rolling back the transactions just like myTransaction.Rollback() in asp.net. But what about exceptions? In asp.net, I am used to doing the following:
<code>Try 'execute commands myTransaction.Commit()Catch ex As Exception Response.Write(ex.Message) myTransaction.Rollback()End Try</code>Will the database inform me of any exceptions (and their messages)? Do I need to put anything explicit in my stored procedure other than rollback transaction?
Any help is greatly appreciated
I have a piece of code that has happily run daily for at least 4 years. I have moved it to a machine that is meant to have an identical set-up, and it fails. I need to work out why it fails, but I am stuck. The code works as follows:
CREATE storedProc1 (Parameter1, etc)
DECLARE @TransactionName nvarchar(127), @TransactionActive bit
-- Plus Other Variables...
BEGIN
BEGIN TRY
SET @TransactionName= 'TransactionName1'
SET @TransactionActive = 'True'
BEGIN TRANSACTION @TransactionName WITH MARK
[Code] .....
When I EXECute StoredProc1, it fails with the Error Message "Cannot roll back . No transaction or savepoint of that name was found." I think that message comes from a ROLLBACK TRANSACTION statement in the CATCH of StoredProc2. I have tried commenting out that ROLLBACK statement, which changes the error message to: "Transaction count after EXECUTE indicates a mismatching number of BEGIN and COMMIT statements. Previous count = 1, current count = 2."
If i want to replicate data only for a particular customer code using SQL 2012 replication - is it possible. I believe that either the entire database or few tables can be replicated using SQL 2012.
View 2 Replies View RelatedI have some problem about Merge Replication.i'm trying to merge Database A to Database B in local Server. So Database i want publisher contain this:
Because only Data in Table change therefore so i choosed it 100% Snapshot was generated after that ( Problem not a valid window user i already figured out )After that. I created Local Subscription in same Server ( Pull subscriptions and Subscription type: Client ),now problem is throw out."The schema script 'vwBuyADT_513.sch' could not be propagated to the subscriber."
I tried research so many time in Google but any information i found isn't useful for this problem.This problem still can ignore,Synchronization still running. But after 4-5 Hours running..This messages throw out: "The Publisher failed to allocate a new set of identity ranges for the subscription. This can occur when a Publisher or a republishing Subscriber has run out of identity ranges to allocate to its own Subscribers or when an identity column data type does not support an additional identity range allocation"
My question is:
Is there anyway solve 2 probem? :
1. "The schema script 'vwBuyADT_513.sch' could not be propagated to the subscriber."
2. "The Publisher failed to allocate a new set of identity ranges for the subscription. This can occur when a Publisher or a republishing Subscriber has run out of identity ranges to allocate to its own Subscribers or when an identity column data type does not support an additional identity range allocation"
We are planning to setup HA using either AAG or FC. ON my production environment (P), I have transnational replication configured for 2 of databases out of three. These data get replicated to another server( C1) hosted on cloud, with local distributor at P.
If I configure AAG/ FCI , how do I handle failover?I wanted to setup P as primary AAG with two replicas as S1 and S2. P->S1 will be synchronous while P->S2 will be asynchronous.Incase my P goes down how the replication will failover S1? Incase P and S1 goes down how do I failover to S2 with the replication.
I wanted to schedule the transaction replication. How do I do it? Currently I have set up a transaction replication which runs continuously and synchronizes the changes with immediate effect.
I need to configure a replication which will gather logs from the publication once in a day.
In ReplMon, some Log Reader agent jobs display as "not running". Rigt-clicking on it the only option is to "Stop" the agent. Seems to me if it's NOT RUNNING you should be able to Start it but that option is disabled. Also, looking at the Agent Job Activity Monitor on the distribution server it says that the job is currently executing.
Running the Replication Agents Checkup job yields nothing and those jobs are still "not running". I can easily Stop/Start the Log Reader in ReplMon or insert a tracer token and then all looks fine. I'm just puzzled by the inconsistency and wondering how I can programmatically check and resolve it.
We are working in a Merge replication environment where we have SQL Server 2005, 11 publications and 2 subscribers.We used to get lot of incidents from the Application owner for blockings, recently we faced a situation where the lead blocker is in sleeping state and the session was used by the merge agent.Checked the query that the session was running, it was sys.sp_MSenumgenerations90;1.
View 1 Replies View RelatedCan we setup a replication ( publisher and subscriber are Sql server) between two different domains, basically different company databases.
View 1 Replies View RelatedI generated a snapshot and it completed Successfully.
How do I apply the snapshot so that the data is refreshed?
[URL]
I have a transactional publication done between two databases on different Virtual machines connected with a network. I noticed that for some reason Every time replication is happening (which is every 2,5 seconds) my data is deleted on the target DB (subscriber) then re inserted.
this should not happen, old data should remain then only the new data should be inserted.
What are the different options for setting up Alerts when Replication stops or errors out?
I ran this code on the publisher:
USE [msdb]
GO
/****** Object: Alert [Replication: agent failure] Script Date: 10/2/2015 11:42:46 AM ******/
EXEC msdb.dbo.sp_add_alert @name=N'Replication: agent failure',
@message_id=14151,
@severity=0,
@enabled=0,
[Code] ....
It already existed.
So I ran it on the Subscriber and it executed.
Are there any third party tools to do this?
We have a publisher sending data to two remote subscribers. Each of these sites is updating a different field in a particular table with its site name and the current date stamp. This data should then sync to each database to show how up to date the last data change was. This lets us keep an eye on whether sync is good or not.
The problem I've got is one subscriber isn't copying its row to the other servers anymore. It gets the row updates from the other sites in the same table but its own updates to this field aren't getting sent across. Nothing shows up in conflict manager for it and nor should it as no other subscriber should be updating this field. If I validate the subscription the field when then get synced but again no updates after the validation will transfer. The other problem which may be related or indicating another issue is the data transfer rate shown in replication monitor is less than 0.1 rows/sec. Reinitializing isn't an option.
So, Microsoft decided that they were deprecating Transactional Replication with Updatable subscriptions. In that case, you have 2 options (if I am correct): Pay for Enterprise (if you are already not) and use peer-to-peer or use bidirectional transactional replication which is basically setting up a transactional from db1 to db2 and also transactional from db2 to db1.
The issue I see in both cases is conflict resolution. With updatable subscriptions, you could specify how to handle the conflict. With either of these 2 options (from what I can tell) you cannot allow the engine to handle this for you.
Any thoughts? Seems like a slap in the face to those who have been using MS for years and a damn good reason for companies that rely on updatable subscriptions to not upgrade to 2012.