You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@activemq.apache.org by davidyce <ye...@ncs.com.sg> on 2016/08/08 04:15:49 UTC
Postgresql Database Lock Issue in Master-Slave setup
Hi,
We have a problem when we shut down serverA (hosting amq master), server B
(hosting amq slave) did not run as Master. In the end, amq becomes blocked.
No incoming, no outgoing messages. Our initial investigation is due to the
amq slave not able to acquire the database lock which amq master is still
holding on.
OS: Centos 7
Activemq version:5.9.0
Postgre database:9.5.3
amq master:61616
amq slave:61617
Appreciate any help.
Thanks in advance.
--
View this message in context: http://activemq.2283324.n4.nabble.com/Postgresql-Database-Lock-Issue-in-Master-Slave-setup-tp4715112.html
Sent from the ActiveMQ - User mailing list archive at Nabble.com.
Re: Postgresql Database Lock Issue in Master-Slave setup
Posted by Tim Bain <tb...@alumni.duke.edu>.
You should test the behavior both when you kill the broker (kill -9 or
equivalent) and when you stop the broker (via the shutdown script). You
may see different behavior between a hard kill and a clean shutdown...
On Aug 10, 2016 7:04 PM, "davidyce" <ye...@ncs.com.sg> wrote:
> hi Tim
> thank you for your response. we tried the setting you recommended and we
> did not face that db lock issue. we are trying out both situations in
> stages first by means of shutting down the entire server. then the second
> stage is by means of killing activemq process. currently we are still at
> stage 1.
> our purpose is to transport messages across 3 tier architecture (web, app,
> database) to ensure data consistency and integrity. the rate of messages
> could be up to 1000 messages per second.
> we are looking forward to activemq and hope we could get continual support.
>
> many thanks
>
> Sent with AquaMail for Android
> http://www.aqua-mail.com
>
> On 9 August 2016 03:08:05 "Tim Bain [via ActiveMQ]" <
> ml-node+s2283324n4715151h9@n4.nabble.com> wrote:
>
> We don't see a lot of traffic on this list about people using PostgreSQL
> (or other SQL data stores, for that matter), and I've seen it said that
> there's been less effort spent to test/maintain/improve the JDBC store
> code, so it's possible that there are bugs that haven't been found yet, and
> this could be one of those.
>
> Can you tell us more about how you shut down the master? Did you shut down
> the ActiveMQ instance itself, or just pull the plug on the server, or...?
>
> And did the slave never successfully start, or just not for some short
> amount of time? The JDBC locker (
> http://activemq.apache.org/pluggable-storage-lockers.html) has a
> lockKeepAlive period; is it possible that you stopped letting the slave try
> to reconnect before that period had elapsed?
>
> Tim
>
> On Sun, Aug 7, 2016 at 10:15 PM, davidyce <[hidden
> email]</user/SendEmail.jtp?type=node&node=4715151&i=0>> wrote:
>
> > Hi,
> >
> > We have a problem when we shut down serverA (hosting amq master), server
> B
> > (hosting amq slave) did not run as Master. In the end, amq becomes
> > blocked.
> > No incoming, no outgoing messages. Our initial investigation is due to
> the
> > amq slave not able to acquire the database lock which amq master is still
> > holding on.
> >
> > OS: Centos 7
> > Activemq version:5.9.0
> > Postgre database:9.5.3
> > amq master:61616
> > amq slave:61617
> >
> > Appreciate any help.
> >
> > Thanks in advance.
> >
> >
> >
> > --
> > View this message in context: http://activemq.2283324.n4.
> > nabble.com/Postgresql-Database-Lock-Issue-in-Master-
> > Slave-setup-tp4715112.html
> > Sent from the ActiveMQ - User mailing list archive at Nabble.com.
> >
>
>
> ________________________________
> If you reply to this email, your message will be added to the discussion
> below:
> http://activemq.2283324.n4.nabble.com/Postgresql-
> Database-Lock-Issue-in-Master-Slave-setup-tp4715112p4715151.html
> To unsubscribe from Postgresql Database Lock Issue in Master-Slave setup,
> click here<http://activemq.2283324.n4.nabble.com/template/
> NamlServlet.jtp?macro=unsubscribe_by_code&node=4715112&code=
> eWVvY2VAbmNzLmNvbS5zZ3w0NzE1MTEyfDI3OTYyNTQ0>.
> NAML<http://activemq.2283324.n4.nabble.com/template/
> NamlServlet.jtp?macro=macro_viewer&id=instant_html%
> 21nabble%3Aemail.naml&base=nabble.naml.namespaces.
> BasicNamespace-nabble.view.web.template.NabbleNamespace-
> nabble.view.web.template.NodeNamespace&breadcrumbs=
> notify_subscribers%21nabble%3Aemail.naml-instant_emails%
> 21nabble%3Aemail.naml-send_instant_email%21nabble%3Aemail.naml>
>
>
>
>
> --
> View this message in context: http://activemq.2283324.n4.
> nabble.com/Postgresql-Database-Lock-Issue-in-Master-
> Slave-setup-tp4715112p4715386.html
> Sent from the ActiveMQ - User mailing list archive at Nabble.com.
Re: Postgresql Database Lock Issue in Master-Slave setup
Posted by davidyce <ye...@ncs.com.sg>.
hi Tim
thank you for your response. we tried the setting you recommended and we did not face that db lock issue. we are trying out both situations in stages first by means of shutting down the entire server. then the second stage is by means of killing activemq process. currently we are still at stage 1.
our purpose is to transport messages across 3 tier architecture (web, app, database) to ensure data consistency and integrity. the rate of messages could be up to 1000 messages per second.
we are looking forward to activemq and hope we could get continual support.
many thanks
Sent with AquaMail for Android
http://www.aqua-mail.com
On 9 August 2016 03:08:05 "Tim Bain [via ActiveMQ]" <ml...@n4.nabble.com> wrote:
We don't see a lot of traffic on this list about people using PostgreSQL
(or other SQL data stores, for that matter), and I've seen it said that
there's been less effort spent to test/maintain/improve the JDBC store
code, so it's possible that there are bugs that haven't been found yet, and
this could be one of those.
Can you tell us more about how you shut down the master? Did you shut down
the ActiveMQ instance itself, or just pull the plug on the server, or...?
And did the slave never successfully start, or just not for some short
amount of time? The JDBC locker (
http://activemq.apache.org/pluggable-storage-lockers.html) has a
lockKeepAlive period; is it possible that you stopped letting the slave try
to reconnect before that period had elapsed?
Tim
On Sun, Aug 7, 2016 at 10:15 PM, davidyce <[hidden email]</user/SendEmail.jtp?type=node&node=4715151&i=0>> wrote:
> Hi,
>
> We have a problem when we shut down serverA (hosting amq master), server B
> (hosting amq slave) did not run as Master. In the end, amq becomes
> blocked.
> No incoming, no outgoing messages. Our initial investigation is due to the
> amq slave not able to acquire the database lock which amq master is still
> holding on.
>
> OS: Centos 7
> Activemq version:5.9.0
> Postgre database:9.5.3
> amq master:61616
> amq slave:61617
>
> Appreciate any help.
>
> Thanks in advance.
>
>
>
> --
> View this message in context: http://activemq.2283324.n4.
> nabble.com/Postgresql-Database-Lock-Issue-in-Master-
> Slave-setup-tp4715112.html
> Sent from the ActiveMQ - User mailing list archive at Nabble.com.
>
________________________________
If you reply to this email, your message will be added to the discussion below:
http://activemq.2283324.n4.nabble.com/Postgresql-Database-Lock-Issue-in-Master-Slave-setup-tp4715112p4715151.html
To unsubscribe from Postgresql Database Lock Issue in Master-Slave setup, click here<http://activemq.2283324.n4.nabble.com/template/NamlServlet.jtp?macro=unsubscribe_by_code&node=4715112&code=eWVvY2VAbmNzLmNvbS5zZ3w0NzE1MTEyfDI3OTYyNTQ0>.
NAML<http://activemq.2283324.n4.nabble.com/template/NamlServlet.jtp?macro=macro_viewer&id=instant_html%21nabble%3Aemail.naml&base=nabble.naml.namespaces.BasicNamespace-nabble.view.web.template.NabbleNamespace-nabble.view.web.template.NodeNamespace&breadcrumbs=notify_subscribers%21nabble%3Aemail.naml-instant_emails%21nabble%3Aemail.naml-send_instant_email%21nabble%3Aemail.naml>
--
View this message in context: http://activemq.2283324.n4.nabble.com/Postgresql-Database-Lock-Issue-in-Master-Slave-setup-tp4715112p4715386.html
Sent from the ActiveMQ - User mailing list archive at Nabble.com.
Re: Postgresql Database Lock Issue in Master-Slave setup
Posted by Tim Bain <tb...@alumni.duke.edu>.
We don't see a lot of traffic on this list about people using PostgreSQL
(or other SQL data stores, for that matter), and I've seen it said that
there's been less effort spent to test/maintain/improve the JDBC store
code, so it's possible that there are bugs that haven't been found yet, and
this could be one of those.
Can you tell us more about how you shut down the master? Did you shut down
the ActiveMQ instance itself, or just pull the plug on the server, or...?
And did the slave never successfully start, or just not for some short
amount of time? The JDBC locker (
http://activemq.apache.org/pluggable-storage-lockers.html) has a
lockKeepAlive period; is it possible that you stopped letting the slave try
to reconnect before that period had elapsed?
Tim
On Sun, Aug 7, 2016 at 10:15 PM, davidyce <ye...@ncs.com.sg> wrote:
> Hi,
>
> We have a problem when we shut down serverA (hosting amq master), server B
> (hosting amq slave) did not run as Master. In the end, amq becomes
> blocked.
> No incoming, no outgoing messages. Our initial investigation is due to the
> amq slave not able to acquire the database lock which amq master is still
> holding on.
>
> OS: Centos 7
> Activemq version:5.9.0
> Postgre database:9.5.3
> amq master:61616
> amq slave:61617
>
> Appreciate any help.
>
> Thanks in advance.
>
>
>
> --
> View this message in context: http://activemq.2283324.n4.
> nabble.com/Postgresql-Database-Lock-Issue-in-Master-
> Slave-setup-tp4715112.html
> Sent from the ActiveMQ - User mailing list archive at Nabble.com.
>