Re: Shared File System Master Slave

[Tim Bain]

Yes, that information is applicable and gives a good overview of the
internals of KahaDB.

Tim

On Mon, Apr 9, 2018, 1:29 AM norinos  wrote:

> I found the following site. Is this helpful for me?
>
> http://www.idevnews.com/images/emailers/110127_ProgressFUSE/WhitePapers/ActiveMQinActionCH05.pdf
>
> --
> Sent from:
> http://activemq.2283324.n4.nabble.com/ActiveMQ-User-f2341805.html
>

Re: Shared File System Master Slave

[norinos]

Hi Tim ! thanks a lot.


> The documentation you linked to is for the AMQ message store, not the 
> KahaDB message store. So if you're using KahaDB as you say, then the page 
> you linked to is irrelevant. 

I finally understood that AMQ and KahaDB are different.

> KahaDB behaves as I described. You only risk losing messages that have not 
> yet been accepted (i.e. you don't risk losing anything). If I recall 
> correctly, the only thing that's flushed to disk periodically is the
> index, 
> and that can be rebuilt from the raw journal files so there's no risk of 
> data loss, just a slower broker startup. 

I feel relieved that it seems there is no risk.


I found the following site. Is this helpful for me?
http://www.idevnews.com/images/emailers/110127_ProgressFUSE/WhitePapers/ActiveMQinActionCH05.pdf





--
Sent from: http://activemq.2283324.n4.nabble.com/ActiveMQ-User-f2341805.html

Re: Shared File System Master Slave

[Tim Bain]

The documentation you linked to is for the AMQ message store, not the
KahaDB message store. So if you're using KahaDB as you say, then the page
you linked to is irrelevant.

KahaDB behaves as I described. You only risk losing messages that have not
yet been accepted (i.e. you don't risk losing anything). If I recall
correctly, the only thing that's flushed to disk periodically is the index,
and that can be rebuilt from the raw journal files so there's no risk of
data loss, just a slower broker startup.

Tim

On Sun, Apr 8, 2018, 10:06 PM norinos  wrote:

> As you say, I'm using KahaDB.
>
> > You asked whether subscriptions would be on the new master, and the
> answer
> > depends on whether the subscription was durable. If it was a durable
> > subscription, the subscription information is persisted and the
> > subscription will be in the same state on the new master, even if the
> > client doesn't immediately reconnect. If the subscription is not durable,
> > then it will be recreated when the client reconnects to the new master,
> so
> > any messages that were unconsumed at the time the original master went
> > down
> > and any messages sent before the client reconnects will be lost for that
> > subscriber.
>
> I'm using durable subscription.
> So, I understand that clients can receive the message without problems
> when switching from master to slave.
>
>
> Please tell me a little more.
>
> ActiveMQ insert cache index into the reference store for performance at
> checkpoint interval.
>
> If master goes down, it seems that chache index can not insert.
> So reference store is not up-to-date.
>
> In this case, if I startup up ActiveMQ, Is the reference store updated at
> the start-up phase?
>
>
>
>
> --
> Sent from:
> http://activemq.2283324.n4.nabble.com/ActiveMQ-User-f2341805.html
>

Re: Shared File System Master Slave

[norinos]

As you say, I'm using KahaDB.

> You asked whether subscriptions would be on the new master, and the answer 
> depends on whether the subscription was durable. If it was a durable 
> subscription, the subscription information is persisted and the 
> subscription will be in the same state on the new master, even if the 
> client doesn't immediately reconnect. If the subscription is not durable, 
> then it will be recreated when the client reconnects to the new master, so 
> any messages that were unconsumed at the time the original master went
> down 
> and any messages sent before the client reconnects will be lost for that 
> subscriber.

I'm using durable subscription.
So, I understand that clients can receive the message without problems
when switching from master to slave.


Please tell me a little more.

ActiveMQ insert cache index into the reference store for performance at
checkpoint interval.

If master goes down, it seems that chache index can not insert.
So reference store is not up-to-date.

In this case, if I startup up ActiveMQ, Is the reference store updated at
the start-up phase?




--
Sent from: http://activemq.2283324.n4.nabble.com/ActiveMQ-User-f2341805.html

Re: Shared File System Master Slave

[Tim Bain]

The AMQ store is not widely used these days. You'll probably be using the
KahaDB store: activemq.apache.org/kahadb.html

As you said, if there is any content that has not been flushed to disk, it
will be lost when the master goes down. However, unless you've manually
set enableJournalDiskSyncs to false, you're not going to lose any messages
whose receipt has been confirmed by the broker. (Any messages lost will be
messages the broker has not yet confirmed receipt of.)

When the slave broker attempts to become active, it will read the data
files and begin serving clients.

You asked whether subscriptions would be on the new master, and the answer
depends on whether the subscription was durable. If it was a durable
subscription, the subscription information is persisted and the
subscription will be in the same state on the new master, even if the
client doesn't immediately reconnect. If the subscription is not durable,
then it will be recreated when the client reconnects to the new master, so
any messages that were unconsumed at the time the original master went down
and any messages sent before the client reconnects will be lost for that
subscriber.

Tim

On Fri, Apr 6, 2018, 1:54 AM norinos  wrote:

> I am planning to use Shared File System Master Slave.(I use AcitveMQ
> 5.13.1)
> http://activemq.apache.org/shared-file-system-master-slave.html
>
> With this system, even though master(#1)'s process goes down due to
> trouble,
> I think that ActiveMQ can be continuously used in slave(#2) process.
>
> I have a question.
>
> http://activemq.apache.org/amq-message-store.html
>
> According to the above documentation, MQ's memory cache periodically
> inserted into the reference store(db.data).
>
> If master(#1) goes down, memory cache cannot insert.
> It seems that the reference sotre index does not match the Data Logs.
>
> At this time, can #2(New master) start normally ?
> If not, What kind of trouble will occur ?
> For example, Subscribed messages will be subscribe agein.
>
>
>
>
>
> --
> Sent from:
> http://activemq.2283324.n4.nabble.com/ActiveMQ-User-f2341805.html
>

Re: Shared File System Master/Slave with DRBD

[Mark Anderson]

Greg,

Thanks for the reply. I just wanted to confirm a couple of things.

1) The NFS requirement is so that DRBD can be used in Primary/Primary mode
as the file system needs to be writeable from both nodes?
2) With NFS I would also need to use Pacemaker to make the NFS server HA?
3) Is the NFS Tech Guide listed at
http://www.linbit.com/en/downloads/tech-guides what you are referring to?

Thanks,
Mark


On 1 May 2014 00:34, Greg Eckert g...@linbit.com wrote:

 Hey Mark,

 You should be able to use drbd with something like NFS without a problem.
 The LINBIT engineers are in the process publishing an NFS Tech guide... For
 current official documentation, feel free to contact me directly.

 Best Regards,

 Greg

 --
 : Greg Eckert
 : g...@linbit.com
 : Not an Engineer
 : LINBIT | Your Way to High Availability
 : Web: http://www.linbit.com

Re: SHARED FILE SYSTEM MASTER/SLAVE (newb)

[Johan Edstrom]

The slave is a standby that will once it can lock the journal reverse the 
roles, so 
the downed node once it is up will start looking for the lock.

On Jan 8, 2013, at 9:38 AM, pico denis.pique...@londonlife.com wrote:

 configured 2 brokers using the shared file syntax
 
 However, the slave broker is always looking for the database connection and
 is not starting up.
 Is this the expected behavior?
 
 Log from the slave broker
 INFO   | jvm 1| 2013/01/08 10:14:40 |  INFO | Database
 \\servername\Shared-Queue\data\kahadb\lock is locked... waiting 10 seconds
 for the database to be unlocked. Reason: java.io.IOException: File
 '\\Ssmsftd1\Shared-Queue\data\kahadb\lock' could not be locked.
 INFO   | jvm 1| 2013/01/08 10:14:50 |  INFO | Database
 \\servername\Shared-Queue\data\kahadb\lock is locked... waiting 10 seconds
 for the database to be unlocked. Reason: java.io.IOException: File
 '\\Ssmsftd1\Shared-Queue\data\kahadb\lock' could not be locked.
 INFO   | jvm 1| 2013/01/08 10:15:00 |  INFO | Database
 \\servername\Shared-Queue\data\kahadb\lock is locked... waiting 10 seconds
 for the database to be unlocked. Reason: java.io.IOException: File
 '\\Ssmsftd1\Shared-Queue\data\kahadb\lock' could not be locked.
 INFO   | jvm 1| 2013/01/08 10:15:10 |  INFO | Database
 \\servername\Shared-Queue\data\kahadb\lock is locked... waiting 10 seconds
 for the database to be unlocked. Reason: java.io.IOException: File
 '\\Ssmsftd1\Shared-Queue\data\kahadb\lock' could not be locked.
 
 
 
 
 
 --
 View this message in context: 
 http://activemq.2283324.n4.nabble.com/SHARED-FILE-SYSTEM-MASTER-SLAVE-newb-tp4661492.html
 Sent from the ActiveMQ - User mailing list archive at Nabble.com.

Re: Shared File System Master Slave Tips?

[Mario Siegenthaler]

We've tried it on NFS on linux and it didn't work (some locking issue
with java and nfs). However it worked with windows shares, but we're
not using it because we run linux boxes.

Mario

On 5/31/07, Michael Slattery [EMAIL PROTECTED] wrote:

Hello List!

I'm interested in putting together the Shared File System Master
Slave scenario in a managed hosting setup.   What are you using to
achieve the shared file system master slave on your installs?

I would appreciate it very much if folks on this list would share
descriptions of successful environments. SAN/NAS, Filesystems,
etc would be very helpful.

If you are doing this in a managed hosting environment, even
better!It's not easy to determine who is doing SAN with shared
access or decent NAS from the vendor websites out there (rackspace
and the like).

Last case, if you tried Shared File System Master Slave but found
that you had to go with JDBC that would also be helpful.

Thank you.
Mike

Re: Shared File System Master Slave with OCFS

[James Strachan]

On 5/19/07, Christopher G. Stach II [EMAIL PROTECTED] wrote:

James Strachan wrote:

 Thanks for the heads up! :)

 I guess we could make the locking strategy pluggable  we could have
 some implementation call the fcntl locking. e.g. maybe using Jtux

 http://www.basepath.com/aup/jtux/


Even though one could achieve this, I don't know what the benefit would
be.


Allow folks to run multiple broker instances on a shared file system
so that if one broker fails, another takes over.
http://activemq.apache.org/shared-file-system-master-slave.html



  It only shifts the responsibility down.  At least the user would
want a broker that is dependent on one or more brokers.  Each of those
brokers shouldn't be dependent on each other for locking, data, or
anything else.  I can appreciate that some people assume that shared
data is available, but shared data is just as easily corrupted, locked,
or unavailable.  Essentially, when you find a single responsibility and
divide it, it probably shouldn't converge somewhere down the line.  This
current pattern is most likely unusable for any HA situation.


Agreed - I'd much prefer a share nothing HA strategy where folks can
spin up and down brokers at any point in time and everything just
works (tm) with messages replicated onto multiple physical stores and
things auto-partitioning etc.

However we're not there yet, so its a question of trade-offs. Some
folks find it easier to rely on their shared file system / SAN to do
the HA; others can use the share-nothing approach in ActiveMQ (Pure
Master Slave)...
http://activemq.apache.org/masterslave.html

and live with the current limitations (no automatic fail-back of the
old master when it is brought back online).

Like many things in IT its a tradeoff. For now, today, the shared file
system approach is a pretty reasonable approach; if you already have a
reliable distributed file system (particularly if you have a SAN).

--
James
---
http://macstrac.blogspot.com/

Re: Shared File System Master Slave with OCFS

[Christopher G. Stach II]

James Strachan wrote:
 
 Thanks for the heads up! :)
 
 I guess we could make the locking strategy pluggable  we could have
 some implementation call the fcntl locking. e.g. maybe using Jtux
 
 http://www.basepath.com/aup/jtux/
 

Even though one could achieve this, I don't know what the benefit would
be.  It only shifts the responsibility down.  At least the user would
want a broker that is dependent on one or more brokers.  Each of those
brokers shouldn't be dependent on each other for locking, data, or
anything else.  I can appreciate that some people assume that shared
data is available, but shared data is just as easily corrupted, locked,
or unavailable.  Essentially, when you find a single responsibility and
divide it, it probably shouldn't converge somewhere down the line.  This
current pattern is most likely unusable for any HA situation.

-- 
Christopher G. Stach II

Re: Shared File System Master Slave with OCFS

[Christopher G. Stach II]

Christopher G. Stach II wrote:
 James Strachan wrote:
 Thanks for the heads up! :)

 I guess we could make the locking strategy pluggable  we could have
 some implementation call the fcntl locking. e.g. maybe using Jtux

 http://www.basepath.com/aup/jtux/

 
 Even though one could achieve this, I don't know what the benefit would
 be.  It only shifts the responsibility down.  At least the user would
 want a broker that is dependent on one or more brokers.  Each of those
 brokers shouldn't be dependent on each other for locking, data, or
 anything else.  I can appreciate that some people assume that shared
 data is available, but shared data is just as easily corrupted, locked,
 or unavailable.  Essentially, when you find a single responsibility and
 divide it, it probably shouldn't converge somewhere down the line.  This
 current pattern is most likely unusable for any HA situation.
 

Wow.  That sounded kind of snotty. :)  I meant client that is
dependent, too.

-- 
Christopher G. Stach II

Re: Shared File System Master Slave with OCFS

[James Strachan]

On 5/18/07, James Strachan [EMAIL PROTECTED] wrote:

On 5/17/07, felipera [EMAIL PROTECTED] wrote:
 Hi everyone,

 I am trying to setup two MQ Servers (4.1.1), sharing the same data directory
 (I tried Derby and Kaha), on top of OCFS, but the locking doesn't seem to be
 working. It works fine when both MQs are running on the same server (still
 using OCFS). I see the second MQ waiting for the lock to be released
 (Journal is locked... waiting 10 seconds for the journal to be unlocked.).
 That's why I am not sure if it's a OCFS issue. But when I run each MQ in
 separate boxes (still sharing the same data directory using OCFS) it doesn't
 work, they both start successfully.

This is the OCFS you're talking about right?
http://en.wikipedia.org/wiki/OCFS2


Actually OCFS2 seems more like a real distributed file system for
general purpose use; the OCFS looks more specifically for using to
host oracle data tables. Am wondering how good the file locking is on
OCFS? Certainly its clear the mutex file locking from Java isn't
supported on OCFS.

--
James
---
http://macstrac.blogspot.com/

Re: Shared File System Master Slave with OCFS

[James Strachan]

On 5/17/07, felipera [EMAIL PROTECTED] wrote:

Hi everyone,

I am trying to setup two MQ Servers (4.1.1), sharing the same data directory
(I tried Derby and Kaha), on top of OCFS, but the locking doesn't seem to be
working. It works fine when both MQs are running on the same server (still
using OCFS). I see the second MQ waiting for the lock to be released
(Journal is locked... waiting 10 seconds for the journal to be unlocked.).
That's why I am not sure if it's a OCFS issue. But when I run each MQ in
separate boxes (still sharing the same data directory using OCFS) it doesn't
work, they both start successfully.


This is the OCFS you're talking about right?
http://en.wikipedia.org/wiki/OCFS2

I guess the file locking isn't working from Java. I wonder how to use
file locking properly on OCFS.


I am thinking about using a network of brokers, since I don't know what else
to try.


If you want clustering  replicated message stores you could try one
of the other Master/Slave implementations.
http://activemq.apache.org/masterslave.html

FWIW Networks of Brokers is for store and forward of messages only;
messages are never replicated in multiple locations with networks of
brokers.

--
James
---
http://macstrac.blogspot.com/

Re: Shared File System Master Slave with OCFS

[Christopher G. Stach II]

James Strachan wrote:
 On 5/18/07, James Strachan [EMAIL PROTECTED] wrote:
 On 5/17/07, felipera [EMAIL PROTECTED] wrote:
  Hi everyone,
 
  I am trying to setup two MQ Servers (4.1.1), sharing the same data
 directory
  (I tried Derby and Kaha), on top of OCFS, but the locking doesn't
 seem to be
  working. It works fine when both MQs are running on the same server
 (still
  using OCFS). I see the second MQ waiting for the lock to be released
  (Journal is locked... waiting 10 seconds for the journal to be
 unlocked.).
  That's why I am not sure if it's a OCFS issue. But when I run each
 MQ in
  separate boxes (still sharing the same data directory using OCFS) it
 doesn't
  work, they both start successfully.

 This is the OCFS you're talking about right?
 http://en.wikipedia.org/wiki/OCFS2
 
 Actually OCFS2 seems more like a real distributed file system for
 general purpose use; the OCFS looks more specifically for using to
 host oracle data tables. Am wondering how good the file locking is on
 OCFS? Certainly its clear the mutex file locking from Java isn't
 supported on OCFS.
 

OCFS2 properly supports POSIX locking semantics with fcntl.  lockf and
flock aren't supported yet.  If that's what the JVM uses under the
covers, you're out of luck.  If this is about OCFS and not OCFS2, I'm
really sorry. :)

-- 
Christopher G. Stach II

Re: Shared File System Master Slave with OCFS

[James Strachan]

On 5/18/07, Christopher G. Stach II [EMAIL PROTECTED] wrote:

James Strachan wrote:
 On 5/18/07, James Strachan [EMAIL PROTECTED] wrote:
 On 5/17/07, felipera [EMAIL PROTECTED] wrote:
  Hi everyone,
 
  I am trying to setup two MQ Servers (4.1.1), sharing the same data
 directory
  (I tried Derby and Kaha), on top of OCFS, but the locking doesn't
 seem to be
  working. It works fine when both MQs are running on the same server
 (still
  using OCFS). I see the second MQ waiting for the lock to be released
  (Journal is locked... waiting 10 seconds for the journal to be
 unlocked.).
  That's why I am not sure if it's a OCFS issue. But when I run each
 MQ in
  separate boxes (still sharing the same data directory using OCFS) it
 doesn't
  work, they both start successfully.

 This is the OCFS you're talking about right?
 http://en.wikipedia.org/wiki/OCFS2

 Actually OCFS2 seems more like a real distributed file system for
 general purpose use; the OCFS looks more specifically for using to
 host oracle data tables. Am wondering how good the file locking is on
 OCFS? Certainly its clear the mutex file locking from Java isn't
 supported on OCFS.


OCFS2 properly supports POSIX locking semantics with fcntl.  lockf and
flock aren't supported yet.  If that's what the JVM uses under the
covers, you're out of luck.  If this is about OCFS and not OCFS2, I'm
really sorry. :)


Thanks for the heads up! :)

I guess we could make the locking strategy pluggable  we could have
some implementation call the fcntl locking. e.g. maybe using Jtux

http://www.basepath.com/aup/jtux/



--
James
---
http://macstrac.blogspot.com/

Re: Re: Shared File System Master Slave with Windows

[Dingwen Yuan]

I have tried the latest version - apache-activemq-4.2-20070328.130210-35, but 
the same error still exists. With what are you going to substitute activeio?

Thank you!




Dingwen Yuan
2007-04-10



发件人： James Strachan
发送时间： 2007-04-03 15:49:52
收件人： users@activemq.apache.org
抄送： 
主题： Re: Shared File System Master Slave with Windows

On 4/3/07, pdvyuan  [EMAIL PROTECTED]  wrote:
 Hi all,

 Can I use a shared windows file system to achieve Shared File System Master 
 Slave? I have made some tests, but occasionally some error would occur. I 
 made the test under 2 instances of activemq-4.2 runing on WinXP. The errors 
 are as follows:

 2007-03-30 17:04:33,640 [28.64.7.49:4858] DEBUG JournalMessageStore   
  - Journalled message add for: 
 ID:xuj01-a7d120-3614-1174637145906-0:0:1:1:296058, at: 39:3829174
 2007-03-30 17:04:33,687 [eckpoint Worker] DEBUG JournalPersistenceAdapter 
  - Checkpoint started.
 2007-03-30 17:04:33,703 [8.64.7.111:1333] DEBUG JournalMessageStore   
  - Journalled message remove for: 
 ID:xuj01-a7d120-3615-1174637145921-1:0:1:1:109282, at: 39:3830090
 2007-03-30 17:04:33,781 [eckpoint Worker] DEBUG JournalPersistenceAdapter 
  - Marking journal at: 39:3829174
 2007-03-30 17:04:33,781 [eckpoint Worker] ERROR JournalPersistenceAdapter 
  - Failed to mark the Journal: 
 org.apache.activeio.journal.InvalidRecordLocationException: The location is 
 less than the last mark.
 org.apache.activeio.journal.InvalidRecordLocationException: The location is 
 less than the last mark.
  at 
 org.apache.activeio.journal.active.JournalImpl.setMark(JournalImpl.java:340)
  at 
 org.apache.activemq.store.journal.JournalPersistenceAdapter.doCheckpoint(JournalPersistenceAdapter.java:400)
  at 
 org.apache.activemq.store.journal.JournalPersistenceAdapter$2.iterate(JournalPersistenceAdapter.java:128)
  at 
 org.apache.activemq.thread.DedicatedTaskRunner.runTask(DedicatedTaskRunner.java:88)
  at 
 org.apache.activemq.thread.DedicatedTaskRunner.access$000(DedicatedTaskRunner.java:25)
  at 
 org.apache.activemq.thread.DedicatedTaskRunner$1.run(DedicatedTaskRunner.java:39)
 2007-03-30 17:04:33,781 [eckpoint Worker] DEBUG JournalPersistenceAdapter 
  - Checkpoint done.

 Is this error due to the insufficient support of exclusive file locks by 
 Windows.

The above looks like a bug. I'm guessing you've used a fairly old
snapshot of 4.2, I'd recommend using a recent one from the end of
March 2007 which no longer uses activeio AFAIK.

http://people.apache.org/repo/m2-snapshot-repository/org/apache/activemq/apache-activemq/4.2-SNAPSHOT/

-- 

James
---
http://radio.weblogs.com/0112098/