cancel
Showing results for 
Search instead for 
Did you mean: 

[Webinar 09/26/2019] EDB Failover Manager - Features and Demo

Highlighted
Community Manager

[Webinar 09/26/2019] EDB Failover Manager - Features and Demo

 

EDB Failover Manager - Features and Demo

 

Date: Thursday, September 26th
Time: 10.00am India | 12.30pm Singapore | 1.30pm Japan | 2.30pm Australia EST | 4.30pm New Zealand

 

Join Postgres expert Saurabh Shelar as he goes through the features and a live demo of EDB Postgres Failover Manager (EFM). You don’t want to miss this!

 

EDB Postgres Failover Manager (EFM) is a high-availability module from EnterpriseDB that enables a Postgres Master node to automatically failover to a Standby node in the event of a

software or hardware failure on the Master.

 

Saurabh will cover key features of EDB Postgres Failover Manager and Demo Failover scenarios.

Highlights include:

  • Tuneable Properties for User Environment
  • Supported Failover Scenarios
  • A step by step demo

 

About the Speaker:

Saurabh Shelar

Saurabh Shelar, Sr. Systems Engineer at EnterpriseDB

 

REGISTER NOW
 
Interested in the webinar but not able to attend the live event? Register anyway and we’ll send you the recording shortly after the webinar.

 

2 REPLIES 2
Community Manager

Re: [Webinar 09/26/2019] EDB Failover Manager - Features and Demo

Follow Up: [WEBINAR]  EDB Failover Manager - Features and Demo

 

We hope you had a chance to join us for the webinar!!   If not, don't worry! 

You can view the webinar via the link below. Be sure to check out the Q & A Session too!

 

 

Q & A SESSION:

__________________

 

Question:

How to decide which node is membership coordinator?

 

Answer:

Whichever is the first node to be added to EFM setup becomes the membership coordinator. If due to some problem, this leaves the EFM cluster. The one added after that assumes the role.

 

_____

 

Question:

When failover happens, is there an easier way to re-instate the original master as slave?

 

Answer:

It will depend on how much time it takes for you to bring the old master back from the outage. if the lag is too much, you might have to recreate it as a slave. EFM creates a recovery.conf file on the master (if the system running the DB itself didn't die) so if the machine is brought online soon it should be able to rejoin without any trouble.

 

_____

 

Question:

We have efm.nodes file which gets overwritten during the agent start stop activity and the entries get cleared. Any suggestions?

 

Answer:

Overwriting can be stopped by setting parametr 'stable.nodes.file' the parameter to 'true'. Details at the following link:
https://www.enterprisedb.com/edb-docs/d/edb-postgres-failover-manager/user-guides/user-guide/3.6/clu...

 

_____

 

Question:

In a 2 node replication setup running on PostgreSQL, how does the EFM licensing work?

 

Answer:

Please contact us at sales@enterprisedb.com

 

_____

 

Question:

Any feature to integrate pgbouncer with EFM? Presently there is no integration and we had to restart pgbouncer after starting EFM services.

 

Answer:

At present pgpbouncer has not been integrated with EFM. However, EFM does provide for calling scripts to automate actions which you might need to be performed after failover. These can be called using 'script.fence' and 'script.post.promotion'. Details can be found at the following link:
https://www.enterprisedb.com/edb-docs/d/edb-postgres-failover-manager/user-guides/user-guide/3.6/clu...

 

_____

 

Question:

In case if master node goes down abruptly recovery.conf is created and database starts in recovery mode. Will the same happen in this version as well or any changes need to be made to bring the setup back intact?

 

Answer:

In case of an outage of the master recovery.conf file will only be created after it has been declared unreachable by the agents based on the timeout parameters specified during configuration and slave will get promoted. If you don't want failover to happen but just get notified about a master going down, you can set 'auto.failover=false'. More on configurable properties at the following link:
https://www.enterprisedb.com/edb-docs/d/edb-postgres-failover-manager/user-guides/user-guide/3.6/clu...

 

_____

 

Question:

What is the difference between setting auto failover in efm.properties to yes and initiating a switchover command in command prompt?

 

Answer:

auto.failover is put into action automatically when EFM detects that the 'Master DB' is unreachable, the prime aim of this is to have a working master. Whereas in switch-over, roles of Master and Slave server are interchanged, here the end result is to ensure that you still have a master and slave arrangement at the end.

 

_____

 

Question:

User EFM is created by default when EFM is installed in RHEL, is it created in this version as well?

 

Answer:

Yes, by default it does create efm user. However, from version 3.0 onwards need for sudoer's permission has been removed. More details at the following link:
https://www.enterprisedb.com/blog/running-edb-postgres-failover-manager-without-sudo

 

_____

 

Question:

Is there a separate license needed for EFM or the Enterprise support includes it?

 

Answer:

Enterprise Postgres enterprise includes EFM:
https://www.enterprisedb.com/pricing#support

 

_____

 

 

 

 

 

 

 

 

 

 

 

Tags (1)
EDB Team Member

Re: [Webinar 09/26/2019] EDB Failover Manager - Features and Demo

Hi,

 

One note about the recovery.conf file that EFM creates. In the case of a master database failure, EFM will create this file to prevent the database from starting. That's to help make it harder for someone to accidentally start it as a seconds master. So the failed master won't come up in recovery -- it won't come up at all.

 

Cheers,

Bobby