About secondary master

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

About secondary master

Hector Zhang
I just found seldom documentation about secondary master. I have following questions:

Should I start the secondary master on the same node of master or another node?
When will it checkpoint the journal, I tried in local mode and found the journal is not changed.
Is there a separate start secondary command? I found that I can only start it on local mode like this: 
bin/alluxio-start.sh local


--
You received this message because you are subscribed to the Google Groups "Alluxio Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.
Reply | Threaded
Open this post in threaded view
|

Re: About secondary master

Andrew Audibert
Hi Hector,

SecondaryMaster exists because Alluxio master can't generate checkpoints while serving requests. This isn't an issue in HA mode since standby masters usually handle checkpointing. But in non-HA mode there aren't any standby masters, so we need a way to generate checkpoints to prevent the journal log from growing unbounded. SecondaryMaster solves this problem. The SecondaryMaster process is a simple process that does nothing except tail the journal logs and generate checkpoints.

There are tradeoffs with running SecondaryMaster on the same node as the primary master.

Same node advantages:
- Doesn't require an additional node
- Can use local filesystem for the journal since it is available to both masters

Different node advantages:
- Each master holds all metadata in memory, so running both masters on the same node doubles the individual node memory requirement.

I would recommend running in HA mode instead of worrying about secondary master in the first place, but if that's not an option I'd say run on the same node unless you expect to have millions of files and limited master memory.

To answer your second question, checkpoints are generated whenever the number of entries since the last checkpoint reaches alluxio.master.journal.checkpoint.period.entries, which defaults to 2 million.

You can explicitly start a secondary master process by running 

bin/alluxio-start.sh secondary_master

- Andrew

On Tue, Jun 26, 2018 at 10:02 PM Hector Zhang <[hidden email]> wrote:
I just found seldom documentation about secondary master. I have following questions:

Should I start the secondary master on the same node of master or another node?
When will it checkpoint the journal, I tried in local mode and found the journal is not changed.
Is there a separate start secondary command? I found that I can only start it on local mode like this: 
bin/alluxio-start.sh local


--
You received this message because you are subscribed to the Google Groups "Alluxio Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.
--

--
You received this message because you are subscribed to the Google Groups "Alluxio Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.
Reply | Threaded
Open this post in threaded view
|

Re: About secondary master

Hector Zhang
Thank you very much for your response.

I get a conclusion:
I can not start secondary master if I were using local filesystem for the journal.
Is it right?

On Friday, June 29, 2018 at 4:16:29 AM UTC+8, Andrew Audibert wrote:
Hi Hector,

SecondaryMaster exists because Alluxio master can't generate checkpoints while serving requests. This isn't an issue in HA mode since standby masters usually handle checkpointing. But in non-HA mode there aren't any standby masters, so we need a way to generate checkpoints to prevent the journal log from growing unbounded. SecondaryMaster solves this problem. The SecondaryMaster process is a simple process that does nothing except tail the journal logs and generate checkpoints.

There are tradeoffs with running SecondaryMaster on the same node as the primary master.

Same node advantages:
- Doesn't require an additional node
- Can use local filesystem for the journal since it is available to both masters

Different node advantages:
- Each master holds all metadata in memory, so running both masters on the same node doubles the individual node memory requirement.

I would recommend running in HA mode instead of worrying about secondary master in the first place, but if that's not an option I'd say run on the same node unless you expect to have millions of files and limited master memory.

To answer your second question, checkpoints are generated whenever the number of entries since the last checkpoint reaches alluxio.master.journal.checkpoint.period.entries, which defaults to 2 million.

You can explicitly start a secondary master process by running 

bin/alluxio-start.sh secondary_master

- Andrew

On Tue, Jun 26, 2018 at 10:02 PM Hector Zhang <<a href="javascript:" target="_blank" gdf-obfuscated-mailto="mpP8Jnh0CAAJ" rel="nofollow" onmousedown="this.href=&#39;javascript:&#39;;return true;" onclick="this.href=&#39;javascript:&#39;;return true;">copperyb...@...> wrote:
I just found seldom documentation about secondary master. I have following questions:

Should I start the secondary master on the same node of master or another node?
When will it checkpoint the journal, I tried in local mode and found the journal is not changed.
Is there a separate start secondary command? I found that I can only start it on local mode like this: 
bin/alluxio-start.sh local


--
You received this message because you are subscribed to the Google Groups "Alluxio Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to <a href="javascript:" target="_blank" gdf-obfuscated-mailto="mpP8Jnh0CAAJ" rel="nofollow" onmousedown="this.href=&#39;javascript:&#39;;return true;" onclick="this.href=&#39;javascript:&#39;;return true;">alluxio-user...@googlegroups.com.
For more options, visit <a href="https://groups.google.com/d/optout" target="_blank" rel="nofollow" onmousedown="this.href=&#39;https://groups.google.com/d/optout&#39;;return true;" onclick="this.href=&#39;https://groups.google.com/d/optout&#39;;return true;">https://groups.google.com/d/optout.
--
Andrew Audibert
<a href="http://alluxio.com/" style="color:rgb(17,85,204);font-size:12.8px" target="_blank" rel="nofollow" onmousedown="this.href=&#39;http://www.google.com/url?q\x3dhttp%3A%2F%2Falluxio.com%2F\x26sa\x3dD\x26sntz\x3d1\x26usg\x3dAFQjCNEOzcgHeqiDCH9tkk9r99TjTZX7Nw&#39;;return true;" onclick="this.href=&#39;http://www.google.com/url?q\x3dhttp%3A%2F%2Falluxio.com%2F\x26sa\x3dD\x26sntz\x3d1\x26usg\x3dAFQjCNEOzcgHeqiDCH9tkk9r99TjTZX7Nw&#39;;return true;">Alluxio, Inc. | <a href="http://bit.ly/alluxio-open-source" style="color:rgb(17,85,204)" target="_blank" rel="nofollow" onmousedown="this.href=&#39;http://www.google.com/url?q\x3dhttp%3A%2F%2Fbit.ly%2Falluxio-open-source\x26sa\x3dD\x26sntz\x3d1\x26usg\x3dAFQjCNEDNVXZleOB7VIXYMM8vGuSeh4NQw&#39;;return true;" onclick="this.href=&#39;http://www.google.com/url?q\x3dhttp%3A%2F%2Fbit.ly%2Falluxio-open-source\x26sa\x3dD\x26sntz\x3d1\x26usg\x3dAFQjCNEDNVXZleOB7VIXYMM8vGuSeh4NQw&#39;;return true;">Alluxio Open Source | <a href="http://bit.ly/alluxio-get-involved" style="color:rgb(17,85,204)" target="_blank" rel="nofollow" onmousedown="this.href=&#39;http://www.google.com/url?q\x3dhttp%3A%2F%2Fbit.ly%2Falluxio-get-involved\x26sa\x3dD\x26sntz\x3d1\x26usg\x3dAFQjCNEMkj0A_5qpmy2ZeIJGUV1QLgzxRg&#39;;return true;" onclick="this.href=&#39;http://www.google.com/url?q\x3dhttp%3A%2F%2Fbit.ly%2Falluxio-get-involved\x26sa\x3dD\x26sntz\x3d1\x26usg\x3dAFQjCNEMkj0A_5qpmy2ZeIJGUV1QLgzxRg&#39;;return true;">Alluxio Community Site

--
You received this message because you are subscribed to the Google Groups "Alluxio Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.
Reply | Threaded
Open this post in threaded view
|

Re: About secondary master

Andrew Audibert
If you are using local filesystem for the journal, secondary master needs to run on the same node as the primary master. Secondary master needs to have access to the journal. So you can't run secondary master on a different node, but you could run in on the same node as the primary master.

On Thu, Jun 28, 2018 at 10:19 PM Hector Zhang <[hidden email]> wrote:
Thank you very much for your response.

I get a conclusion:
I can not start secondary master if I were using local filesystem for the journal.
Is it right?


On Friday, June 29, 2018 at 4:16:29 AM UTC+8, Andrew Audibert wrote:
Hi Hector,

SecondaryMaster exists because Alluxio master can't generate checkpoints while serving requests. This isn't an issue in HA mode since standby masters usually handle checkpointing. But in non-HA mode there aren't any standby masters, so we need a way to generate checkpoints to prevent the journal log from growing unbounded. SecondaryMaster solves this problem. The SecondaryMaster process is a simple process that does nothing except tail the journal logs and generate checkpoints.

There are tradeoffs with running SecondaryMaster on the same node as the primary master.

Same node advantages:
- Doesn't require an additional node
- Can use local filesystem for the journal since it is available to both masters

Different node advantages:
- Each master holds all metadata in memory, so running both masters on the same node doubles the individual node memory requirement.

I would recommend running in HA mode instead of worrying about secondary master in the first place, but if that's not an option I'd say run on the same node unless you expect to have millions of files and limited master memory.

To answer your second question, checkpoints are generated whenever the number of entries since the last checkpoint reaches alluxio.master.journal.checkpoint.period.entries, which defaults to 2 million.

You can explicitly start a secondary master process by running 

bin/alluxio-start.sh secondary_master

- Andrew

On Tue, Jun 26, 2018 at 10:02 PM Hector Zhang <[hidden email]> wrote:
I just found seldom documentation about secondary master. I have following questions:

Should I start the secondary master on the same node of master or another node?
When will it checkpoint the journal, I tried in local mode and found the journal is not changed.
Is there a separate start secondary command? I found that I can only start it on local mode like this: 
bin/alluxio-start.sh local


--
You received this message because you are subscribed to the Google Groups "Alluxio Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].

For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "Alluxio Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.
--

--
You received this message because you are subscribed to the Google Groups "Alluxio Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.