flink-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Conrad Crampton <conrad.cramp...@SecData.com>
Subject Re: Reload DistributedCache file?
Date Thu, 17 Aug 2017 08:00:55 GMT
Whilst this looks great and thanks for pointing this out, the problem isn’t so much as with
identifying changes in the HDFS file as the overhead in re-loading it isn’t significant
(10s of lines long) but the problem is more to do with using the Flink mechanisms i.e. DistributedCache
and refreshing this on a local level (to the taskmanager node).

I may have to read HDFS directly in the FilterFunction to keep it up to date.

Thanks
Conrad

From: Ted Yu <yuzhihong@gmail.com>
Date: Wednesday, 16 August 2017 at 18:11
To: Conrad Crampton <conrad.crampton@SecData.com>
Cc: "user@flink.apache.org" <user@flink.apache.org>
Subject: Re: Reload DistributedCache file?

For hdfs, there is iNotify mechanism.

https://issues.apache.org/jira/browse/HDFS-6634

https://www.slideshare.net/Hadoop_Summit/keep-me-in-the-loop-inotify-in-hdfs

FYI

On Wed, Aug 16, 2017 at 9:41 AM, Conrad Crampton <conrad.crampton@secdata.com<mailto:conrad.crampton@secdata.com>>
wrote:
Hi,
I have a simple text file that is stored in HDFS which I use in a RichFilterFunction by way
of DistributedCache file. The file is externally edited periodically to have other lines added
to it. My FilterFunction also implements Runnable whose run method is run as a scheduleAtFixedRate
method of ScheduledExectutorService which reloads the file and stores the results in a List
in the Filter class.

I have realized the errors of my ways as the file that is reloaded is the cached file that
is copied to temporary file location on the node which this instance of Filter class is loaded
and not the file from HDFS directly (as this has been copied when the Flink job started.

Can anyone suggest a solution to this? It is I think a similar problem that Add Side Inputs
in Flink [1] proposal is trying to address but not finalized yet.
Can anyone see a problem if I have a thread that reloads the HDFS file being in the main body
of my Flink program and registers the cache file within that reload process e.g.

env.registerCachedFile(properties.getProperty("whitelist.location"), WHITELIST);

i.e. does this actually copy the file again from HDFS to temporary files on each node? I think
I’d have to have the same schedule I have currently that reload within my Filter function
too though as all the previous process would do is to push the HDFS file to temp location
and not actually refresh my List.

Any suggestions would be welcome.

Thanks
Conrad

[1] https://docs.google.com/document/d/1hIgxi2Zchww_5fWUHLoYiXwSBXjv-M5eOv-MKQYN3m4/edit#heading=h.pqg5z6g0mjm7

SecureData, combating cyber threats

________________________________

The information contained in this message or any of its attachments may be privileged and
confidential and intended for the exclusive use of the intended recipient. If you are not
the intended recipient any disclosure, reproduction, distribution or other dissemination or
use of this communications is strictly prohibited. The views expressed in this email are those
of the individual and not necessarily of SecureData Europe Ltd. Any prices quoted are only
valid if followed up by a formal written quote.

SecureData Europe Limited. Registered in England & Wales 04365896. Registered Address:
SecureData House, Hermitage Court, Hermitage Lane, Maidstone, Kent, ME16 9NT



***This email originated outside SecureData***

Click here<https://www.mailcontrol.com/sr/MY4bnrNl+sPGX2PQPOmvUrxsj2JZCuwu2DTYNQkvjbg!dlAMiOPe32HUzm7icNaCC2UNWDX6TnbJaydNu2tBwg==>
to report this email as spam.
Mime
View raw message