Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [geomesa-users] age-off strategy, is there one?

Andrew,

 

That timeline is fine.  I think we do not have a very complicated age-off scenario at the moment, just date is fine.  So if it takes about a week or so for us to have a solution (and understand how to apply it and what it may mean) that allows us to age-off that would be appreciated!

 

Thanks,

Diane

 

From: geomesa-users-bounces@xxxxxxxxxxxxxxxx [mailto:geomesa-users-bounces@xxxxxxxxxxxxxxxx] On Behalf Of Andrew Hulbert
Sent: Friday, April 28, 2017 11:30 AM
To: geomesa-users@xxxxxxxxxxxxxxxx
Subject: Re: [geomesa-users] age-off strategy, is there one?

 

Yes, I think that I probably need to do a little more testing really quick to verify that it is working properly before you use it. You have to be careful with age off iterators because it's easy to misconfigure if you're not careful :)

We will merge it into the 1.2.x codebase and master soon. What's your time frame? Is a week or so ok?

Also, what is your age off strategy? So this repo: https://github.com/jahhulbert-ccri/geomesa-ageoff-itr is actually a copy that I made of another repository. The other one is the one I use in production is more specific to a deploy environment and also uses data-derived authorizations and retention information based on customer business rules. Do you need to do anything like this? E.g. multiple kinds of data come in with different tags and a date. The tags define the retention periods and the iterator can smartly delete them based on a map of tag -> retention.

Andrew


On 04/28/2017 11:17 AM, Diane Griffith wrote:

Andrew,

 

We are on 1.2.7.2 currently.  So one could merge that iterator over and then just apply the age off filter based on the DTG field specified as the time field on ingest?

 

Any idea if there are plans to add this type of iterator to the main code base?  I have to see if a custom code build is amenable or not so that is one area of concern for me but we do have to age off the data.

 

Thanks,

Diane

 

From: geomesa-users-bounces@xxxxxxxxxxxxxxxx [mailto:geomesa-users-bounces@xxxxxxxxxxxxxxxx] On Behalf Of Andrew Hulbert
Sent: Friday, April 28, 2017 11:14 AM
To: Geomesa User discussions
Subject: Re: [geomesa-users] age-off strategy, is there one?

 

Diane,

I do have a GeomesaAgeOffIterator that I use in production but it is not merged into the main codebase. The version I have works on 1.2.x at the moment. You are on 1.2.7.3 right?

You can age things off with delete commands manually but this is very tedious. I think the iterator is a better way. It applies at scan, minc, and majc and can actually handle things other than just dates. (e.g. data auths)

I'll send you something off list.

Andrew

On 04/28/2017 11:04 AM, Diane Griffith wrote:

Is there a recommended strategy one can use to age-off data leveraging accumulo age-off that will clean up data after a certain date.  I did not think just applying an age-off filter manually would work.

 

We are currently on the 1.2.x code base but I did see there was a GeomesaAgeoffIterator.scala so is this something more readily supported in the 1.3.x code base?

 

Is there a more manual recommended way to do age-off (either define a date in the SFT that can be used) across all tables created?  If there is not one the only thing we can figure which we want to avoid is querying and doing manual deletes and assume that will clean up all tables for a given sft (indexes and all).

 

Thanks,

Diane





_______________________________________________
geomesa-users mailing list
geomesa-users@xxxxxxxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://dev.locationtech.org/mailman/listinfo/geomesa-users

 




_______________________________________________
geomesa-users mailing list
geomesa-users@xxxxxxxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://dev.locationtech.org/mailman/listinfo/geomesa-users

 


Back to the top