Tag Info

New answers tagged

0

It's unclear what they want to measure. To see the network lag, all they need is ping in the OS. That should be much less than a second, even to the other side of the world. (Unless they have a sloppy network.) Seconds_behind_master is problematical. It is a handy tool for the DBA, but it can be "wrong" for many reasons. When running a long query, it ...


0

You could make server2 as slave of server1. Server2 contains a replica of db1. Into db2 create a view that point to db1.t1, and int db3 create a view that point to db1.t2. I have attached an image to clarify.


0

It seems sensible to me that they would want to check the lag between the master and slave. It's one of the key things we monitor here, as if the slaves start getting behind it could be a sign of a problem, it also reduces their usefulness for reads / backups etc. The easiest way would be for them to run SHOW SLAVE STATUS on the Slave and look at the ...


2

To initialize a Transactional subscription from a backup you actually end up creating the publication first, enabling it to initialize subscriptions from a backup, then take a backup, and then initialize subscriptions with the backup. Changes that occur after taking the backup and before initializing the subscription(s) will be stored in the distribution ...


1

I have personally had experience with Transactional Replication. In my case, I configured the Distributor at the Report Server, and then created the Publication on Live server to be controlled from the Report Server. The Subscription from the Report Server worked perfectly. With this configuration, I was successfully able to replicate data from the Live ...


0

If I understand correctly: You have two DB's in a Master <-> Master setup, but one of them has the replicate-do-db rules meaning they are now (potentially) out of sync? At present the DB with the replicate-do-db options is only replicating (processing) statements where the database concerned is in the list (the precise rules vary depending on whether you ...


1

DISCLAIMER : Not a Neo4j expert IMHO I would yes because the architecture link you gave sounds a lot like Percona XtraDB Cluster. The underlying replication scheme is driven by Galera. In essence, all nodes check with one another before doing transaction commits or rollbacks. This also means that you can run inserts, updates, and deletes on any node in ...


0

This is a known issue with SQL Server 2012. It used to work in previous versions. The Microsoft Connect item can be found here: https://connect.microsoft.com/SQLServer/feedback/details/769502/permissions-not-replicating-for-certain-object-types-when-using-a-sql-server-2012-distributor We first noticed it on the MSDN Forums here: ...


1

The Snapshot Agent process account used to connect to the Distributor must be db_owner in the distribution database and have read, write, and modify permissions on the snapshot share. The account used to connect to the Publisher must be db_owner in the publication database. The Log Reader Agent process account used to connect to the Distributor must be ...


0

Since the mysql schema is usually manipulated with GRANT and REVOKE commands, there is no need to let the mysql schema replicate. Why ? Someone could hack privileges in by means of INSERT commands into mysql.user. Such hacks are possible because I wrote about it before Oct 28, 2011 : MySQL: trying to create root user Sep 12, 2011 : How do you gain access ...


0

You are on the right track, but let me refine it STEP 01: On Slave1, run STOP SLAVE; STEP 02: On Slave1, run SHOW SLAVE STATUS\G until Seconds_Behind_Master is 0 STEP 03: On Slave1, write down Relay_Master_Log_File (RMLF) and Exec_Master_Log_Pos (EMLP) from STEP 02 STEP 04: On Slave2, run SHOW SLAVE STATUS\G until Seconds_Behind_Master is 0 STEP 05: On ...


0

There is some steps missing, depending on your needs. I am assuming you are using traditional, non-GTID replication. Before doing any topology change, make sure that your slaves have the exact same dataset than the master. The current setup, with 2 jumps is double the change of having suffered some replication woe and having drift. I wouldn't go into ...


0

Add additonal step to avoid any confusion you could do flush tables with read lock after replication is caught up as it will do global lock and would avoid conflict in case of any change coming up, use unlock tables at the end to release the lock acquired. Moreover it would be better to have this done in low/off production time in case it is in production, ...


3

Having done this type of project before, here's some of the practical gotchas to look out for: "2. This master then does one-way slave propagation to up to 200 read-only servers via VPN." Think carefully about the network traffic required here. In the simplest terms, if the insert/update/delete load on the database takes up 1/200th of a network cable's ...


0

The important parts of the query look like SELECT p...., a...., r.... FROM tblPriceForAllotment p JOIN tblAllotment a ON p.AllotRecID = a.RecID JOIN tblProduct r ON r.ProdCode = p.ProdCode WHERE p.ProdCode in(... long list ...) AND p.PriceDate>='2015-10-09' AND p.PriceDate<='2015-10-17' ORDER BY ProdCode; ...


0

I used @RolandoMySQLDBA's suggestion and commented out slave_parallel_workers, which defaults to 0, effectively disabling multi-threaded replication. This fixed the issue, which was addressed in the MySQL 5.6.23 release notes.


0

Yes Relay slaves are possible. But let's try to speed up the queries first. The PROCESSLIST smells like you are using MyISAM, and that switching to InnoDB might help a lot. Please show us the rest of select p.ApproxPrice1, p.ApproxPrice2, p.ApproxPrice3, p.ApproxPrice4, p.PromoMessage, p.NightsToSta, together with EXPLAIN EXTENDED SELECT... and SHOW ...


0

Guess what ? You have a nasty little bug (Replication stall with multi-threaded replication) This bug was first reported in MySQL 5.6.17 on 20 Jun 2014 15:45 SUGGESTION #1 You need to look over all the release notes from MySQL 5.6.18 up to MySQL 5.6.22 to see if it was resolved. If any of those release notes claim the problem was fixed, then someone ...


1

Correct, skipping skips the transaction not the statement and that is why it is better to investigate the failure than blindly skipping. You're best placed to use pt-table-checksum and pt-table-sync to resolve drift after a skip since you will likely have diverged.


0

I found a simple solution here. Distribution Agent - Start /Stop in T-SQL -- To STOP the Distribution Agent: sp_MSstopdistribution_agent @publisher, @publisher_db, @publication, @subscriber, @subscriber_db --To START the Distribution Agent: sp_MSstartdistribution_agent @publisher, @publisher_db, @publication, @subscriber, @subscriber_db My complete ...


2

Oracle Golden Gate has always supported replication between Oracle and SQL Server. For a project I recently confirmed that Golden Gate for Oracle 10 is still available if you push a bit. If you are using an Oracle version that is out of support then you need to update for a number of reasons. Synchronization with Azure is detailed here.


-1

I apologise if I misunderstood (I don't use Amazon RDS). But can you use: SET SESSION SQL_LOG_BIN = 0; -- DO STUFF HERE SET SESSION SQL_LOG_BIN = 1; After setting it to 0, it will prevent everything in your current session being written to the Binary logs. Setting it back to 1 will resume writing to be binary logs. Everything else continues as per ...


0

ANSWERS TO YOUR QUESTIONS 1) Since I have a Master/Slave structure, do I need to manage this and change the Slave also? Or any structural changes will directly affect the Slave? I would recommend you change the Master to all InnoDB and leave the Slave all MyISAM You will definitely need to make configuration changes since InnoDB has more a lot more ...


0

I can suggest a crude method that worked effectively for me . Create a python script that polls a particular table after some milliseconds and prints the time stamp when the count of rows changes. Keep this script running on both master and slave . So when you insert a record on the master you will get the timestamps when the record becomes available on ...


0

Are you starting PostgreSQL on the standby using the postgresql command ? If yes then you should use the pg_ctl command as it detaches the DB server process from the terminal. With the postgresql command it will only display the log output on the screen like you have described and stay there because it is a standby server.


0

I'm guessing your log on server A is bieng shared by the two databases (master A and the slave of B) It looks like you have set ssl = on in your master's postgresql.conf file and have set ssl off in your slave(B). You should either set ssl on on all servers or none. PostgreSQL doesn't usually have a problem replicating across minor versions


1

A smaller shell script can help the cause assuming you have catered all the differences between master/slave.. provided you have login path created if not use -uroot -ppass otherwise. Moreover you may need to check output log and need to correct the out of sync tables manually. c=mysql --login-path=lgp --protocol=TCP -NBe "show status like ...


1

Assuming you have corrected the slave drift, you can just "start slave" again, but You may need to skip some transactions, if they have been manually applied to the slave- you can do that with SET GLOBAL sql_slave_skip_counter You probably have other differences between the master and the slave, because of the operations you mentioned or because of the ...


0

What you did is correct as per the procedure and messages found under the mongod log is something expected and will display unless you perform operations with a mongo shell and it should allow you to connect to localhost or any remote server and then to the desired database. Since you had cancelled the running processes with Ctrl-C event, it had been ...


1

If the binlog file number reaches its maximum value, it appears the server will throw an error. However, the maximum value is not 999999 as it would appear ... it's 2,147,483,647. From the source file sql/log.h (MySQL 5.6 source tree): #define MAX_LOG_UNIQUE_FN_EXT 0x7FFFFFFF Other relevant files are sql/binlog.cc and sql/log.cc. The code that ...


2

Yes, it is very possible. As the Documentation says, you should have a minimum of three nodes in a Percona XtraDB Cluster. You only need one node out of the three nodes. Make sure all nodes in the PXC have unique server_ids. Make sure the Master you are replicating from has a unique server_id from the PXC nodes Make sure the master and every PXC node is ...


1

There are cases in Merge Replication when parent and child rows can be processed out of order on synchronization resulting in constraint violations and conflicts. To understand how and why this can occur and how it may apply to your specific scenario, I recommend reading through: How to understand Merge Replication article processing order Merge Article ...


1

I am adding an additional answer because Giovanni's one is completely right, but it only defines why scaling, how to do HA on the master and on the slaves, and how to split read queries among the slaves from an architectural point of view. I think OP question was more about how to do the read-write split itself. Then answer depends on the software stack ...


1

A good solution is to define two datasources for the applications. One datasource pointing to master is for writing. One datasouce pointing to slave node is for reading. The applications must be aware that data can be delayed with respect to the master. There are applications that could tolerate a slight delay on data, and other applications that must be ...


1

1) Yes 2) it's not going to be a problem but of course it depends on the problem itself. Lets use the term - incident. Problem is already a problem by definition. Solve incidents, analyze the cause , prevent the problem!!! ;) 3) in case of an error, try to think what caused it. Nothing to worry about specifically in my opinion Try to strictly separate ...


1

"Is it possible to do so? The trigger will insert some rows into a unique table that exists only on the slave." yes, you can A lot of warehousing architectures are based on the slave only. The title of your question is wrong. Change replica DB with SLAVE ONLY


0

It looks like you asked 2 questions: Question 1 Does this mean I should go with multi-master solution if I want to promote standby to accept write transaction as well? Built-in PostgreSQL features are enough to get a WAL-based, read-only standby (a.k.a. secondary, a.k.a. slave) server. However, they are not enough to get multi-master operation. ...


0

i use http://en.wikipedia.org/wiki/Corosync_%28project%29 and peacemaker. maybe http://clusterlabs.org/wiki/PgSQL_Replicated_Cluster can help you.



Top 50 recent answers are included