Apache Cassandra Cassandra Community Webinar Series, 6 Oct 2014

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Webinar | Things You Should Be Doing When Using Cassandra Drivers

Description

Did you know there are some things you should doing when writing your data driven application using Apache Cassandra? Let’s talk about that. There are features in almost every driver that will keep your application online and running fast. You should know what they are. Even if you know these features, I’m going to tell why they work and why they’re a good idea. This will not be language specific; I will be using multiple languages and drivers. This talk should appeal to programmers in general.

A

Everyone, this is Rebecca mills, I'm, a junior evangelist with data specs today I'm here to tell you about what I think is a really important topic and it's a topic that isn't really known a lot of great detail by a lot of people.

A

So these are things you should be doing when using costanzo drivers, regardless of what programming language you're using and now I, don't seem to be an expert, but I have done. Some investigating and I picked out some important feature that I think everyone should be using to make their experience using the Cassandra drivers a good one.

A

So, a little bit a little about what sorry one second bring some technical difficulties here, a.

A

Little about what I do at the company all right one. Second, there.

B

You go a little.

A

About what I did the company I try to create a create awareness for open source, Cassandra and I? Do that by developing content like blog post, Orioles, walkthroughs and webinars, and I also have my own Twitter account, which you guys should definitely check out. Part of my job is also to step into the shoes with movie and try to determine what kind of problems people just being introduced to Cassandra might be encountering, which may not be so obvious to an expert.

A

Then this whole newcomer prospective extends to the drivers as well. I want to make the driver seemed, is easy and approachable as I can for everyone, so I'm developing beginners beginners content around them as well.

A

So this webinar is not language specific, I'm going to jump around a couple languages in my examples like Python and Java, but since the AP is, are pretty much the same across all the data stacks drivers, almost everything I talk about is applicable to any of them.

A

The dataset revised drivers for Java Python and C sharp and most recently Ruby nodejs in C++, so that's really exciting. There's also loads of open-source drivers to choose from so you have a lot of options.

A

We have links to a lot these on the planet, Cassandra site. I would actually think of it as the definitive list there's also some tutorials I've done up. You can have a look at.

A

Okay, so, first of all look at some really basic common sense kind of things we should be doing when using a driver. So if you're really new to the drivers, this is kind of bread and butter.

A

So a cluster innocence allows you to configure different important aspects of the way, connections and queries really handled at this level. You can configure contact point, which is the addresses of the nodes to be contacted initially before the driver performs no discovery, there's also retry policies and load balancing policies.

A

Usually such settings are done once at the application level.

A

While the API the session is centered around query execution, it also does some heavy lifting behind the scenes, as it manages turno connection rules synth of the long-lived object. It should not be used in a short list fashion with requests and responses you're going to want to use the same cluster and fashion instances across your application.

B

B

A

This is what should be at the top of your application code and only one time use the session in the rest of your code.

A

So if you want to execute a statement more than once, you should be used, you should be using prepared statements, it's the better alternative to having data. You want inserted hard-coded. It is part of a string. In this case. Our query is only sent over the server in the power Swan. We just change the parameters each additional time.

A

We will talk about this a little more depth than a few seconds.

A

Let's get into some cool as well as useful features, you should be using to have the best programming experience with Cassandra, regardless of your programming language. Now, I want to look at those features you should be using if you want your client to be interacting with Cassandra and the most performant way possible. These features should definitely speed things up a bit I.

A

Actually look cool and useful to gather myself.

A

We're going to look at for paired statements, load, balancing policies and as a part of that retry policies in connection pooling, as well as a think API.

A

Overall, prepare statements are more performance in using string.

A

They will be parsed only once on the server side as a poster straighten, which will be parsed every time it's sent.

A

So, if you're using the same query over and over with different parameters, this is definitely the way to go.

A

Also because parameter is in our heart code into the string, but since separately we are able to avoid sequel injection.

A

So, let's consider a string.

A

Our data for Bob Jones is baking to the string we stand over to the server. This is one of we are only sitting one row of data into our database.

A

A

But we might want to add multiple users with the same insert statement. We might actually want to add hundreds or thousands of users, it's practical to use if it's impractical to use strings. At this point, you.

A

A

Now is prepared statements today, I want to settle along with the string. This gets fed separately as a parameter almost like a variable. The prepared statement only gives two at once, and the parameters are the only thing that changes.

A

So what's the difference behind the scene.

A

As opposed to sending your queries at string, we which results in a large amount of data at a large power, cause this prepared statement. They only get parts on the server side once they are cached on the server and concurrent hashmap stored in memory. The next time. Only the parameters and a prepared statement IDR sent over the wire. The md5 party prepared statement is a key matching, a value in a hashmap.

A

If we send our insert statement with our and bits our data embedded part of the string, it's a large amount of data in a large parts cost when every single string must be parsed over the server. So we can see why a prepared statement was that we be a better option.

A

So what does this all means? The developer?

A

It means that your queries are much faster and York. Why is much more performance?

A

This is a blog post from that this is from a blog post at Netflix shown the difference in through in throughput when using prepare statement versus not you can see there's about five times as much through flip when using prepared statements.

A

Probably was very obvious to everybody: is that using prepared statement of for loop is just a big no-no. This would cancel out all the performance benefits of using a prepared statement in the first place. The query is built-in slant again each time and thus river ends up being bombarded with the same query over and over. This is probably even less efficient than using string you so now we're going to talk a little bit about load, balancing load. Balancing policies are responsible for routing requests.

A

The term ring which knows the driver must connect to, as well as the order in which different nose will be tried in case of network failures. Each cluster can be configured with a specific, no balancing policy to be used and that same policy will be used for all requests: the costs all sessions managed by the cluster instance.

A

So we're going to look at a little example, as we conceded clusters configured with a total where policy wrapped around a DC aware round-robin policies and we're also using a default retry policy.

A

We're also just going to use a simple example. Data model we've created a user's table with a user name, first name and last name. We're also going to insert two rows of data to get ourselves started.

A

Our cluster for discussion purposes has a replication factor of three simplify our token to simplify our token ranges from zero to a hundred, but real token ranges are actually much bigger. Each node is responsible for one. Fourth of the token range clients the client connects to the node listed in the contact points, the mid-penn dot, 0, dot, 0 dot. One is aware of the entire cluster and token layout and the no respond with a list.

A

The list and populates the populate, the connection pool and each token range stored per node, including replicas. When the client execute the Select statement, the statement is parsed. Will this be with a consistency level of one? For this example, the primary key is examined and the string are Mills is hashed using murmur three, the string, our meals hashes to token 15.

A

Super load balance since our load balancing policy is token aware. We will now use the token produced by the hash by hashing the primary key using the token 16. We will then look at for the corresponding nodes.

A

The token 15 matches on range is held by three nose.

A

Those three nodes will now be addressed when looking for token 16, we also added a data center away around Robin policy.

A

This means we will first look at only the replicas in the local data center, which is DC one annotated here.

A

Round robin is a method that cycles through the list of replica nodes. For that token range, we will start at the first node listed, then the second node and then the third note, and it repeats.

A

It will skip any non replica nose in the connection pool.

A

Now, a situation where a note is down, we start at the first node as usual, and then the second.

A

Once we get a timeout, the note is marked down from the connection pool retreat. We retry the same request on the third replica and it repeats.

A

In this case, we're going to eliminate the token aware, routing and just use round-robin.

A

We're splitting a look at the token string, p mcfadden is hash using murmur three and we see the string mcfadden hashes the token 77.

A

In this case, we don't use a token for load balancing decisions, a string, p mcfadden is only on three nodes in a list.

A

That is only one hop away. On the first note.

A

First, two notes on the third note: since the bat is not on this node, we need to coordinate to a replica that has been sent back to the coordinator and the round robin continues.

A

Trencher, adding one more hop to coordination, it can be inefficient, especially as the cluster gets larger.

A

So the default behavior used to be round robin. This is change due to the fact that token aware is much more performant. It acts as a filter around a child policy, and only replicas will be considered when using it, so we won't be wasting our time on non replica knows waiting for them to coordinate.

A

I also wanted to mention the whitelist policy in passing. This policy ensures that queries are only performed on a given list of nodes of your choosing and they limit the effect of auto-discovery of pure nodes.

A

The native by memory primary, the native binary protocol supports pipeline, which means multiple requests and responses can be sent over the same connection simultaneously.

A

This makes asynchronous statements possible.

A

We don't have time to wait. We don't have to wait for a response from our cluster to continue running our code. Instead, almost immediately, we look at a future object, which is like a promise. The future is a placeholder for the statement that was just executed. I also want to mention in passing that all asynchronous methods and then a sink.

A

When you do a select query and execute a sink, the future that holds the execute for you all, you do. The future holds that executes for you. While you do other work, this execute gets run on the server. When you do ask the result, it's in blocks it's useful of doing concurrent coding and, if used properly, it can be really fast session. Got execute, just sits there. Otherwise.

A

You think futures can be handy if you're executing things over and over again, you can send them down the pipe and when you iterate over these futures, you can pull out the results as they come in.

A

So now that you've heard about all these really cool features with the drivers, you may be wondering where you can download them for yourself and give them a test run.

A

A

The clinic offender is a great place for clothes and resources. You can find all sorts of stuff there, like blog posts, webinars tutorials and a whole lot more, but it's also a great place for all your driver needs so once you're on planet Cassandra, you're going to want to go to the download section and into the client to drop first section. I.

A

Spend a good bit of time delving into the different drivers and I've develops in the walkthroughs, taking you through the basics of getting up and running with the drivers, there's even a lot of topics that I did not touch on in this webinar.

A

Here you can find links to other the data sections as well as well as many of the open-source drivers.

A

So if you guys are just getting started with Apache Cassandra or any of the other drivers, I would love to hear about your experience. You can easily find me on Twitter. Thank you guys for tuning in to my webinar and if there's any questions, you can go ahead and ask them now.

B

Alright, well so we have openings for questions and answers, Thank, You Rebecca. So let me ask you this real quick too. So right now we we just released a ruby driver and an oj s driver.

B

So you've been writing a lot of content right, but this stuff yeah.

A

B

Planet is Android arm yeah.

A

When I Cassandra, where you can find a lot of the tutorials I've, got off beginner tutorials for different datastax drivers, mm-hmm I recently just put out one for the new Ruby driver. So.

B

Now, with those I mean it does, it require a lot of experience with Cassandra or how can someone get started with Cassandra too well.

A

I would suggest going actually you you need the key space and the setup from the try Cassandra section of the website, so I would recommend someone go through the try to stand through a tutorial part of the website and then move.

B

A

Drivers I right.

B

We have a few questions, actually quite a few so and thank you for those. So the first question here is a prepared statement bound on one session or is it usable on another session. You.

A

Can only only found on one session, you can't use it on another one. So.

B

I would have to if I wanted to use a same type of prayer or two different sessions. What would.

A

Make any sense, would it no no.

B

You wouldn't have two sessions running at the same time, right even.

A

All these so here one key space against.

B

Someone they were trying to make to two different clusters would have to have two different sessions: okay,.

A

B

A

B

Don't know when you would want to do that, but interesting questions I'd point out it like you said sessions are there's one session for code of the right right.

A

B

So you, you would put everything into that session. Yeah.

A

B

Okay, so uh how about the right? So this is another question here: does the query builder in Java have a way for generating prepared statements so.

B

So that we didn't really cover query builder yeah.

A

That's actually, I did a part of one of my tutorials. I did up on Panik Sandra Dee was about query builder, but that well.

B

In a nutshell, what I mean? Let's we don't need to go into big detail here, but I mean there's. Probably a lot of people are like weird. Those are you didn't talk about clear in a nutshell, what is query builder? What does it do? I mean? Is it? Is it just for building the query period.

A

It's only cuz like that yeah, it's only in the Java driver, yeah.

B

Okay, only it's actually.

A

Don't like the way to get around sequel injection so.

B

It's only in the java Turkey, that's probably worth pointing out to is that you know it's not.

A

B

To all datastax drivers, but.

A

B

Think I think the question is whether it generates a prepared statement that is like a prepared statement when you use query bills.

A

Yeah essentially the same thing: it.

B

A

Ward using strings that's another way of avoiding to young strapping right.

B

In my mind, there is eventually.

A

Yeah prepared statement and of itself mm-hmm.

B

Right, so how do you get the token range it's using the Cassandra driver? Oh man, that's in an advanced user here.

A

You can't get the total range is using a Cassandra driver. Unfortunately, no.

B

But um I believe that is a that's a feature. Elfman right token, the token command, I believe yeah.

A

I think you can get them.

B

A

Command through the is a component of sikhi, well yeah.

B

So you'd have to have dude capsize use a tow good.

A

B

You seek you out, yeah.

A

Dude, how do you seek, will teach you well or to discover so contagious? It.

B

May be and but might be worth looking into, but is not an easy one to answer now. Maybe that's somewhere underneath you can ask me, but it's a murmur three hatch right. Yes,.

A

B

You can just hash it yourself and get to token actually that's easier day. That's.

A

Probably the easier way to do it right, I.

B

Think we just answer that question so for good token ranges maybe put it for the individual servers as well. So um actually I have a more advanced answer that you look at the system's table in Cassandra. The token ranges are in there, so that's more secret. So what is the best practice for writing? In batches Oh someone's been watching my Twitter stream I.

B

Mean I I think we're asking her. How would you do batches? Oh.

A

It's just small batches yeah.

B

A

Recommend yeah yeah.

B

And then um that's kind of near and dear to my heart right, because I actually just filed a jira on friday too. Well, there's a warning jura for large batches. Now, there's a drought for failing on large batches bachelors are a pretty much. A large batches are an anti-pattern right.

B

Yeah I would like to say that reserved for atomicity and not an optimization, do not do use large batches. Are you going to hit my own personal, pet, peeve Bachelet? If I see you large batches going into the system, I I could I don't get on you about it next question. Could I see an example? One more time of the future results that array for execute in a sink I. Think you're going to have your slides online yeah.

A

I'll put those up phone slideshare right, yeah.

B

Plana Cassandra yeah it'll be on getting.

A

Up there, some somewhere.

B

A

They'll be out there.

B

Yeah yeah, so we can answer it. That way. Is there a convenient way of mapping between futures return for the a thing calls and scallops future classes? We don't.

A

B

A skylit drive away yeah.

A

We don't really.

B

Have a skydiver at.

A

This point yeah, like well I, mean if that.

B

That I think that something's being talked about right now.

A

Yeah, we're kind of in the process of embracing skele, more right.

B

I mean, but we saw that was a Python example. You did yeah.

A

And yes, it was.

B

My thumb in the java java de teachers as well that's.

A

B

A

Really and rubia yeah.

B

Route will rubies are promises, yeah, so I think I think it would be natural. The thing to assume that spells future classes would map as well. uh Here's an easy one for me: can we get this recording sure we'll be putting it online soon? Just watch Rebecca's, a washer or Twitter and she'll be posting. Whenever this is online or planets Sandra we do post our webinars out, probably two or three days after we're done. It'll.

A

B

Week all right: how about batches spanning a key big boy, people love their batches, don't they.

B

I mean but I think the question is: can you in a single bat, can you do.

A

More than one cases, yes, you can it's.

B

Just a bunch of different season right.

A

B

Now do you gain right performance by using a token aware connection I think we saw that graph.

A

Right, yeah yeah, because you're not wasting any time on knows that or not. Oh, we.

B

Saw the one a prepared statement: I'm sorry, I kind of threw me out there yeah.

A

B

Are token aware, we never saw a graph for that. No.

A

B

Think it might be some good content.

A

B

But I mean that that's! Why we're using toko, where the race for perform.

A

That's exactly what we're using to open, where yeah.

B

But Stefan I think you actually opened up the door for a really great blog post about what whether or not token were and how efficient it really is. But I we.

A

Do know that it is.

B

More efficient because a good measurement be good. um Can you query the local table I'm, not sure what that means. We vocal table, um maybe system table.

A

I'm not really sure what the question Jeff.

B

Yeah, if you have a clarification, went to type to invite the team asked where, when do you set the consistency level in the Java driver that just set for the one statement or on the session yeah.

A

You set the consistency level on the whole session. Once is the whole application.

B

Then I, so that means, when you see when you create the session, there's a because it simply level like a default consistency, level, yeah.

A

B

And then of what about, if I want to do with / connection, yeah.

A

You can also do it for connection if you wanted to well.

B

A

B

Also, a retry policy that changes, consistency at level 2 right, yeah,.

A

There's also downgrading consistency, level, retry policy, where, if you have trouble connecting such that, you can't get meet your consistent level requirements it'll actually down that take to bring the consistency level down such that on your next reach. How you look at a response, but you won't be reaching the consistency level you would set so.

B

That would take like the about the quorum and go down to oh yeah.

A

Depends on how big your customers yeah.

B

Tim watch out for that one yeah.

A

B

A consultant for years I always worried when people change their consistency level just to try to get away everything, but it might make sense in some cases right. Yeah.

A

Depends on how rigid you want to be about your consistency, level right and the.

B

Rule still stands right, every ever requested to Cassandra kid have a consistency level in it. Yeah.

A

B

Sort yeah so so air gap so to confirm the load. Balancing the client can be aware, which knows, are the primary replica notes so that you can skip the notes that don't hold data if the ring is out of whack and needs to be repaired, you'd have to scan the entire ring big questions. Let's break this dish.

A

B

The client is a whiz, so it's a client aware of which one which knows have the data. For the token.

A

Yes, if you're using token aware and.

B

So yeah I think that's the core yeah.

A

B

So basically you're asking it still gonna we're saying absolutely yes,.

A

B

Then, though, this is, if the rings out of whack it needs to be repaired, you'd have to scan each high rank, I mean I, don't think that doesn't really matter in this case, I mean your data from there because it needs to be repaired. That's not going to be that's not going to be in a table, though, we're not going to see like repair status and a connection pool lottery. No.

A

I, wouldn't think so, though,.

B

And so in that case, yeah I think it's just if your your clusters, that out of whack I, think you got another problem here, not some new drivers really going to help you with, in this case right mmhmm,.

A

B

But if I think I, if I get you right here, eric is you're worried about this inconsistency, because your data is really last up in your ring fix your ring. First, the drivers not going based on that one.

B

Large data sets using spring batch. Okay, that's that's I! Think.

A

B

Question here we go again: hey Rebecca can I do really large batches I, wouldn't.

A

Recommend it now: that's.

B

An anti-pattern.

A

B

It seriously is the default threshold for warning on batch sizes. 5K, it's a little sighs, don't do large batches batches as an old relational guy. I can tell you I used to use a lot of asset there, because it makes sense because it was on a single server, don't use large batches and cassandra, because what you're doing is you're essentially turning the coordinator node into a big load balancer and it just is very inefficient. You're better using I mean there's our buddy right, execute a sink right.

A

You get ya, go.

B

Through all the different keys and just executed a thin protocols, very nice fat and it's much more efficient, alright batches are my pet peeve and what can be considered a large batch I just said that I mean that was 5k in total size is where you start seeing warnings in the log. You know fight these batches of more than 100, then I guess start getting nervous. Like a hundred statement. It's just not efficient you're, better off doing what like one prepared statement right, we.

A

B

A sink sessions are long live instances. However, session has closed method. Should you one be closing a session ever yeah.

A

You should close it's good practice. The closure session once you're.

B

Done running yeah that actually that's a good point to make a it's like leaving a silo yeah.

A

B

You wouldn't I mean sure you could do exit and just shut the whole thing down, but it will stay open, Brandon ass. What details do future results that produce for like record info success, fail reason or fail, I mean so. Basically, what is the future results that the.

A

Result sense, yeah.

B

Yeah, so it is I mean it's just a placeholder right.

A

B

It's a response, I guess, but it's just a result set though, but assuming that now whyn't, so the future isn't blocking. But when does it block blossom.

A

You finally call a future right.

B

Yeah, okay, all right! So there you go Brandon! Guess it's just a result set. So whatever information is inside the results? That will be a part of that, but.

A

B

You're going to block with a smart, sorry ass. What are about the spark? Oh here we go spark right on because standard connector good night, but I use it as a skala driver. I might have been mistaken. Let me take down because actually, this is on the spark side. um No.

A

B

Spark connector is actually just for spark which interesting? Is there a lot of things in there that make it look like spark? It looks like a spark driver, but it's not directly. The state route I think that's coming.

A

B

um You have slides available online, we already.

A

Have available yeah and.

B

Shortly, can you comment on in versus multiple parallel queries, so this is I, think there's a pretty common question, so an in statement, meaning like where you say we're in and where primary key is in. It just creates its list.

B

Yeah I mean I wouldn't use that. Would you know.

A

B

And then, but we just commented on this, multiple parallel queries like async doing a thing stuff. I think that's that's what I think. That's what we're asking here, Jeff right so I, would say definitely using multiple Tyrell of kerla queries where it's at yeah.

A

B

Just based on what you showed right that with the async protocol, the fact that it's going to use all the different round-robin token aware everything each.

A

B

Connection is going to be much more efficient that way a prepared statement prepared on one session should have to be replicated prepared on a different session in order to be used in different sets of batches right. Oh man.

A

B

Prepared one session: okay, positional data matches again here we are with the different session I'm, not sure why you would want more than one session if you're connected.

A

B

Multiple clusters, perhaps that.

A

Would be the only circumstance mm-hmm.

B

A

Just stay close.

B

A

B

And that goes to the the need to have the word of what's going on actually behind the scenes with a session right, I mean it's connecting to the cluster and I was connection pool right.

A

It manages the copra note, connection, pool yeah, so.

B

If you would really need.

A

No, you should need multiple sessions right.

B

A

Batches across right.

B

Jeff system dot local for tokens. Yes, you can actually do that way. That's what we were mentioned before is go into the system table and you can.

A

See the damages and.

B

Again system table is completely open to speak. You out, you can do a select office system tables mm-hmm. So in regards to the nodejs driver, the current one dot 0 beta drivers not backward compatible with 12 I noticed the docs backward compatible. Slider window and I saw that to do comments in the code was the roadmap on releasing one dot. Oh good, you know, I haven't.

A

Had a good look at San ojs driver yet which, where it just came out.

B

A

B

Moving pretty fast yeah.

A

B

Would expect it? That's probably gonna be the next few weeks yeah when died to compatibilities a tougher, because that's different well binary protocol yeah.

A

B

So it's just a different backward compatibility for the binary protocol, but I it. It is coming pretty quick and you could probably pose a question inside the JIRA to do or there's the github issues. If you were looking for it exact date, but I know it is moving rather quickly. I've talked with that team. Quite a bit, they're very excited about that driver and there's a lot of excitement around ojs yeah.

A

There seems to be yeah.

B

And Ruby I was actually really surprised. We've been talking to a lot of people are using their Ruby driver now, yeah.

A

There's a lot of people feeling pretty excited about that one mm-hmm.

B

A

B

Fast, um so what are the difference between cql based drivers and thrift based API is over here we go see two elders of thrift. How do we determine which one should be used in what scenario? So when.

A

Would I ever use the rift.

B

A

Would never use strict anymore.

B

A

The way to go I mean.

B

We were, we've been talking to people at oracle, openworld a lot of experience, dql people a lot of experience through people, but I think that the overriding thing is just the the future of it right. I mean Kristin 30 is is frozen.

A

Yes, I'm frozen yeah.

B

A

Going forward with cql yeah.

B

I think it uh really is all about the feature set. Yes,.

A

B

You're doing any new development: this is what you're doing on all your pages now yeah.

A

I'm only using CQ all these drivers, it's.

B

Really the way forward, yeah.

A

B

Yeah drift drift is just it's an older way of getting things done enough good sport duck. So Jeff is back on this week. You could, you would connect to the cause to retrieve the nodes. We create new connections for each cluster, using the whitelist one with the node.

A

B

Query: the system, local cable, obtaining token, sorry hope that makes sense. No Jeff, you make no sense at all I'm kidding, um yeah, so I, trying to parse out why you're doing this connected to the cluster. True, who knows, create new connections each cluster you're still connecting one cluster you're not going to find in may be that your looks like you might be. Switching cluster and knows is that terminology a cluster is going to be all the nodes in your system even across multiple data centers.

B

So the only reason you would connect with different cluster is, if you've had specifically created a new cluster for some reason: I, where.

A

I've never seen that in a.

B

Production environment, where you have multiple clusters connective as possible yeah, but I Jeff, you might wanna- might as want to hit us up on Twitter and get a little bit or maybe stack overflow and kind of get into your use case a bit because it might be something we need to dig into yeah yeah.

A

B

Close a grad that should be a good students of mentioned to sakura, which, what's a great.

A

Yeah so the great place, if you have any questions a lot of people out there, probably healthier you're.

B

On there, yes, I mother, prepared prepared team does not perform well in a loop. What is the best way for to batch multiple inserts through the prepared statement? I think he was pointing out your example where you did a prepared statement inside.

A

B

What was the better one with that I know, that's what you do right, I mean: where would you put the prepared statement and aleut? Oh it.

A

Would be at the where you do the binding with the parameters yeah.

B

I mean I think that's what about Kane yeah.

A

B

Yeah I mean that's doing a prepared statement once I mean I use. In your examples you you can create like a static in the beginning, a final for the prepared statement in the beginning of your files right.

A

B

You just do it one time sandy and then inside the loop.

A

In the list would be where you do this.

B

A

You find the parameters and.

B

Going back to your diagram, it's what happens every time, I hate that loot.

A

B

Mean it's just sending over what does this thing know, which is observers the with.

A

The prepared statement ID reversal.

B

Yeah, along with the.

A

Parameters right.

B

So that is ND dats, that that.

A

Would be the portion loop right.

B

Should that we cassander 1 dot, 3 I, don't think so we have to it. There is going to be no 13 at 30. There's a 13 secure protocol. Maybe okay, there. It is. He dead, basically, problems between one dot, three binary protocol on a 2 0 briner. Yes, of course, that that's what we were mentioning earlier is that we there are different binary protocols that, from when we went from 12 to 1 20 mm-hmm.

A

B

Ability issues can a think API be used with the.net framework, 40 or higher versions. hmm You know much about that particular angle and tushar baird. We need to.

A

Get Luke yeah that one that gave.

B

Of I'm going to ask you to do me a favor here we so one of the evangelists on our team name is Luke. Tillman he's on Twitter I. Think you'd probably be the best person to ask for that. These are he worked a lot with C sharp edge? He works with Windows Azure. He will be able to answer that question hit him up on Twitter only said hi, but I think that he would probably be the best person to answer that gave.

A

Alls West's load, balancing policy from client driver.

B

Matter if the partition key is used in all queries,.

A

Yeah that should be a way around. It actually use having to use it. If you do specify the partition, Cavin well,.

B

That and that's what that's, what it tokenize is right.

A

Yeah exactly yeah.

B

So I mean low balancing policy with a token we're kind of requires. It I mean.

A

B

Gonna have to the dot. In.Net is a Cassandra road. I'll get value of origin.

A

B

Is gonna wear the only way to retrieve value to reach Colin, it seems so much slow, Dave IM, to tell you we probably are not the best personality. That's people to ask this question and it again.

A

B

Like to hit up Luke I think you could probably answer a lot of questions about this. For sure, is there any support for retry policy with a back off and I? Think that would be this I'll? The downgrading.

A

Consistency: uh okay,.

B

All right, maybe yeah so yeah yeah.

A

B

That's that I think that's what we're talking about for back off or it could be that the back off would be. It would wait for a few seconds and then reconnect, maybe potentially.

A

B

A

Doesn't exist but.

B

You could write it mm-hmm.

A

B

That so that's an interface right. The back of the return policy yeah.

A

It's an interface so.

B

You could implement your own type of retries.

A

Yeah you can customize it to your own one. If you wanted yeah.

B

A

B

The desolation yeah so yeah Gary it that's actually a good point to make- is that if you do have some custom requirements on any of these policies, it's an interface you can you.

A

Can you can tweak it to whatever way works best for you, mmhmm.

B

Yeah and downgrading downgrading retry policies, not one best way, no.

A

B

Tomas says hello:.

A

B

Difficulties of paging through a rosette and went through the training, a Java development could figure out some examples. How do they how to do it? The right way, there's some example available on datastax, web or elsewhere. Do we have a keychain example?

B

You know I, don't know. Datastax examples is a github repo, okay, I believe there is an example on there for paging, okay,.

A

B

That's actually pretty new, so Tomas. If you look on github, there's a there's, a a user called data, stacked examples check that out. There's a lot of good code they're actually for everybody on this I'll call that I think there was actually a directly a paging example and how to do paging with and again that's what version tru dot. Oh right.

A

B

Yeah to do it above has built in Beijing sport Brandon asked when opening a new connection should I use one host, IP or multiple well,.

A

If you use one host IP autodiscover the a little connect to the entire cluster I, don't really I think.

B

It goes along line with seed nodes. If you can do one yeah.

A

B

Having more than one gives you a little more resilience, yeah.

A

That one out disappears.

B

A

B

And we want to have a situation where you're relying on just one note for everything and it's just better off to have more than one too just in case. The co structured DC is a constructor for DC, aware. Round-Robin policy is to find when using wit, load-balancing policy.

B

A

In my mind, is.

B

B

I, don't know depends on that, should be there and it is there I believe that is because you're using not using the current driver, that's the only thing I can think of that's, probably the one dot to driver not to 20.

A

B

He is that used to use a will, look at the policy default or something like that at what point does a parallel do parallel statements performs that are in in the in clause? I mean that time, let me and did there a limit on how many key should put any in clause I mean so basically I I think the first question is: are there? Are there parallel statements and in clause no.

A

There isn't parallel statements in their costs, I.

B

Mean that that scheming goes directly to the coordinator, yeah.

A

B

Yet it in clauses are, are you do not get parsed by the driver as actually was it I think it was a feature request, but as of now, they don't get parsed on it on the client driver, so that just gets sent to the coordinator and then the coordinators, it's up to the coordinator to deal with it and what's the limit now, there's a trivia question: what is the limit? I think I've heard this writing at 64,000.

A

B

Yeah your question I.

A

B

You can the most you can put into it, but it's one of those practical limits, I think Jeffrey, that if you are putting in a lot of values in your in-laws you're going to see a performance hit, it's just like doing a large batch. It's going to hurt bust up I. Actually you know we were just talking about this. Instead of using an in-use, prepares to payment and execute a salute, yeah.

A

That's probably a better idea: yeah.

B

Take advantage of futures get that nice parallelism concurrent program.

A

B

We end up having to cash too many prepared siemens across sessions. I, don't know if that's a question, let.

A

Me go yeah. Do we have a default.

B

Retrying for connections on a read/write operation and.

A

Yeah yeah the default will cover the read and write operations recovering it and are there.

B

Per host at for connections in Cassandra, of course, yes,.

A

B

A

B

It did that's, I mean if you're impatient all about it. Gmx would be Java right. So if you're.

A

B

You can also you can ask for statistics, I mean if you're using Java driver. I use metrics the metrics library for that next question. How do I get a list of nodes, / keys base.

B

So I think that really that commit which node pretty sweet, though.

A

B

Really it doesn't make any sense, actually so the key space, it's not a key space, would be a cluster yeah.

A

I just have notes for key spaces. Yeah.

B

A

Would actually be.

B

A list of notes per cluster requestor is the entity right.

A

B

Spaces inside the cluster so.

B

So I, but getting a list of notes to fit the matter of you can go and ask for that right. So you can say good just about local. Yes,.

B

So, okay bachelors the 5,000 5,000 level, refer to the data size or number two is sighs its its size. So when we're talking 55 k that total size of the bachelor self, so if you had three statements that were really big then and they go over, 5k is really a matter of managing him out. Jdm feat that you're throwing at this thing it's a warning: it's not going to stop it, but it's just.

B

You should be paying attention two batches that have large size like that: the guidance less than hundred statements less than 5k data, this general rule of thumb and it's King one or the other.

B

So hundred statements is quite a few, but then, if it's less than 5k, not as big of a deal, if you did the 5k limit, no matter how many statements you have still one, you need to think about you're much better off doing concurrent programming, no problem all right last couple of questions here and we're going to wrap up what should what should we do to avoid timeouts Cassandra connection from low use application having to reconnect every five minutes?

B

Well bar / club? Sorry, if this is the Java driver, we're using where you won't have to set right, it's just going to do a persistent connection. Yeah.

A

It should just stay connected excusing.

B

A

B

That so I don't it. You know, I hear this question sometimes from Oracle developers, and this is something that I am near and dear to my heart, with Oracle you had so much setup costs and every connection that you know you would connect to it. It would take forever to log in get all the stuff that just doesn't exist in Cassandra.

B

Alright, do you suggest a special driver property to use a huge data load like loading, a few million rows of a thousand bytes each? uh He wants to write a little.

B

So you suggest using a driver property.

A

B

I think is lee exactly this: is there any tweaks to make larger a larger set like that work? No.

A

B

A

There is any yeah.

B

I mean for that I mean it's a few million rows of a Meg each I. It's.

A

Nothing I mean there's nothing to drivers.

B

Will get help you down there yeah it's more of a probably of more of a server consideration? Okay! Well, I think this wraps it up. Thank you for all the really interesting questions. Again, you know check out planet, Cassandra, dot, org and the try Cassandra and look at the client drivers pages. There's a lot of tutorials on their questions. You can always go to stack overflow, we're always there Twitter they def Rebecca myself.

B

We mentioned we threw Luke in there Luke Tillman and if you want to do some free training and I also recommend that Academy datastax calm good for you good for me, and it is one hundred percent free, and that is a good thing for your career. Believe me, we've been doing a lot of work on that lately, so please got something there for you. Thank you very much and we'll see you at the next webinar.