I'm a beginner to DynamoDB, and my online constructor doesn't answer his Q/A lol, and i've been confused about this.
I know that the partition key decides the partition in which the item will be placed.
I also know that the number of partitions is calculated based on throughput or storage using the famous formulas
So let's say a table has user_id as its partition Key, with 200 user_ids. Does that automatically mean that we have 200 partitions? If so, why didn't we calculate the no. of partitions based on the famous formulas?
Thanks
Let's establish 2 things.
A DynamoDB partition can support 3000 read operations and 1000 write operations. It keeps a divider between read and write ops so they do not interfere with each other. If you had a table that was configured to support 18000 reads and 6000 writes, you'd have at least 12 partition, but probably a few more for some head room.
A provisioned capacity table has 1 partition by default, but an on-demand partition has 4 partitions by default.
So, to answer your question directly. Just because you have 200 items, does not mean you have 200 partitions. It is very possible for those 200 items to be in just one partition if your table was in provisioned capacity mode. If the configuration of the table changes or it takes on more traffic, those items might move around to new partitions.
There are a few distinct times where DynamoDB will add partitions.
When partitions grow in storage size larger than 10GB. DynamoDB might see that you are taking on data and try to do this proactively, but 10GB is the cut off.
When your table needs to support more operations per second that it is currently doing. This can happen manually because you configured your table to support 20,000 reads/sec where before I only supported 2000. DynamoDB would have to add partitions and move data to be able to handle that 20,000 reads/sec. Or is can happen automatically to add partitions because you configured floor and ceiling values in DynamoDB auto-scaling and DynamoDB senses your ops/sec is climbing and will therefore adjust the number of partitions in response to capacity exceptions.
Your table is in on-demand capacity mode and DynamoDB attempts to automatically keep 2x your previous high water mark of capacity. For example, say your table just reached 10,000 RCU for the first time. DynamoDB would see that is past your previous high water mark and start adding more partitions as it tries to keep 2x the capacity at the ready in case you peak up again like you just did.
DynamoDB is actively monitoring your table and if it sees one or more items are particularly being hit hard (hot keys), are in the same partition and this might create a hot partition. If that is happening, DynamoDB might split the table to help isolate those items and prevent or fix a hot partition situation.
There are one or two other more rare edge cases, but you'd likely be talking to AWS Support if you encountered this.
Note: Once DynamoDB creates partitions, the number of partitions never shrinks and this is ok. Throughput dilution is no longer a thing in DynamoDB.
The partition key value is hashed to determine the actual partition to place the data item into.
Thus the number of distinct partition key values has zero affect on the number of physical partitions.
The only things that affect the physical number of partitions are RCUs/WCUs (throughput) and the amount of data stored.
Nbr Partions Pt = RCU/3000 + WCU/1000
Nbr Partions Ps = GB/10
Unless one of the above is more than 1.0, there will likely only be a single partition. But I'm sure the split happens as you approach the limits, when exactly is something only AWS knows.
Related
I have an AWS application where DynamoDB is used for most data storage and it works well for most cases. I would like to ask you about one particular case where I feel DynamoDB might not be the best option.
There is a simple table with customers. Each customer can collect virtual coins so each customer has a balance attribute. The balance is managed by 3rd party service keeping up-to-date value and the balance attribute in my table is just a cached version of it. The 3rd party service requires its own id of the customer as an input so customers table contains also this externalId attribute which is used to query the balance.
I need to run the following process once per day:
Update the balance attribute for all customers in a database.
Find all customers with the balance greater than some specified constant value. They need to be sorted by the balance.
Perform some processing for all of the customers - the processing must be performed in proper order - starting from the customer with the greatest balance in descending order (by balance).
Question: which database is the most suitable one for this use case?
My analysis:
In terms of costs it looks to be quite similar, i.e. paying for Compute Units in case of DynamoDB vs paying for hours of micro instances in case of RDS. Not sure though if micro RDS instance is enough for this purpose - I'm going to check it but I guess it should be enough.
In terms of performance - I'm not sure here. It's something I will need to check but wanted to ask you here beforehand. Some analysis from my side:
It involves two scan operations in the case of DynamoDB which
looks like something I really don't want to have. The first scan can be limited to externalId attribute, then balances are queried from 3rd party service and updated in the table. The second scan requires a range key defined for balance attribute to return customers sorted by the balance.
I'm not convinced that any kind of indexes can help here. Basically, there won't be too many read operations of the balance - sometimes it will need to be queried for a single customer using its primary key. The number of reads won't be much greater than number of writes so indexes may slow the process down.
Additional assumptions in case they matter:
There are ca. 500 000 customers in the database, the average size of a single customer is 200 bytes. So the total size of the customers in the database is 100 MB.
I need to repeat step 1 from the above procedure (update the balance of all customers) several times during the day (ca. 20-30 times per day) but the necessity to retrieve sorted data is only once per day.
There is only one application (and one instance of the application) performing the above procedure. Besides that, I need to handle simple CRUD which can read/update other attributes of the customers.
I think people are overly afraid of DynamoDB scan operations. They're bad if used for regular queries but for once-in-a-while bulk operations they're not so bad.
How much does it cost to scan a 100 MB table? That's 25,000 4KB blocks. If doing eventually consistent that's 12,250 read units. If we assume the cost is $0.25 per million (On Demand mode) that's 12,250/1,000,000*$0.25 = $0.003 per full table scan. Want to do it 30 times per day? Costs you less than a dime a day.
The thing to consider is the cost of updating every item in the database. That's 500,000 write units, which if in On Demand at $1.25 per million will be about $0.63 per full table update.
If you can go Provisioned for that duration it'll be cheaper.
Regarding performance, DynamoDB can scan a full table faster than any server-oriented database, because it's supported by potentially thousands of back-end servers operating in parallel. For example, you can do a parallel scan with up to a million segments, each with a client thread reading data in 1 MB chunks. If you write a single-threaded client doing a scan it won't be as fast. It's definitely possible to scan slowly, but it's also possible to scan at speeds that seem ludicrous.
If your table is 100 MB, was created in On Demand mode, has never hit a high water mark to auto-increase capacity (just the starter capacity), and you use a multi-threaded pull with 4+ segments, I predict you'll be done in low single digit seconds.
I have a stream of data, containing a key, that I need to mix and match with data associated with that key. Each key belongs to a partition, and each partition can be loaded from a database.
Data is quite big and only a few hundred out of hundreds of thousands partitions can fit in a single task manager.
My current approach is to use partitionCustom based on the key.partition and cache the partition data inside a RichMapFunction to mix and match without reloading the data of the partitions multiple times.
When the number of message rate on a same partition gets too high, I hit a hot-spot/performance bottleneck.
What tools do I have in Flink to improve the throughput in this case?
Are there ways to customize the scheduling and to optimize the job placements based on setup time on the machines, and maximum processing time history?
It sounds like (a) your DB-based data is also partitioned, and (b) you have skew in your keys, where one partition gets a lot more keys than other partitions.
Assuming the above is correct, and you've done code profiling on your "mix and match" code to make that reasonably efficient, then you're left with manual optimizations. For example, if you know that keys in partition X are much more common, you can put all of those keys in one partition, and then distribute the remaining keys amongst the other partitions.
Another approach is to add a "batcher" operator, which puts up to N keys for the same partition into a group (typically this also needs a timeout to flush, so data doesn't get stuck). If you can batch enough keys, then it might not be so bad to load the DB data on demand for the partition associated with each batch of keys.
Does DynamoDB scale when partition key bucket gets huge? What are the possible solutions if this is really a problem?
TL;DR - Yes, DynamoDB scales very well as long as you use the service as intended!
As a user of Dynamo your responsibility is to choose a partition key (and range key combination if need be) that provides a relatively uniform distribution of the data and to allocate enough read and write capacity for your use case. If you do use a range key, this means you should aim to have approximately the same number of elements for each of the partition key values in your table.
As long as you follow this rule, Dynamo will scale very well. Even when you hit the size limit for a partition, Dynamo will automatically split the data in the original partition into two equal sized partitions which will each receive about half the data (again - as long as you did a good job of choosing the partition key and range key). This is very well explained in the Dynamo DB documentation
Of course, as your table grows and you get more and more partitions, you will have to allocate more and more read and write capacity to ensure enough is provisioned to sustain all partitions of your table. The capacity is equally distributed to all partitions in your table (except for splits - though, again, if the distribution is uniform even splits will receive capacity uniformly).
For reference, see also: how partitions work.
We have a table which is of size 100TB and we have multiple customers using the same table (i.e every customer uses different where conditions). Now the problem statement is every time a customer tries to query the table it gets scanned from top to bottom.
This creates lot of slowness for all the queries. We cannot even partition/bucket the table basing on any business keys. Can someone provide solution or point to similar problem statements and their resolution.
you can provide your suggestions as well as alternative technologies so that we can pick the best suitable one. Thanks.
My 2 cents: experiment with an ORC table with GZip compression (default) and clever partitioning / ordering...
every SELECT that uses a partition key in its WHERE clause will
do "partition pruning" and thus avoid to scan everything [OK, OK, you said you had no good candidate in your specific case, but in general it can be done so I had to mention it first]
then within each ORC file in scope, the min/max counters will be
checked for "stripe pruning", limiting the I/O further
With clever partitioning & clever ordering of the data at INSERT time, using the most-frequent filters, the pruning can be quite efficient.
Then you can look into optimizations such as using a non-default ORC stripe size, a non-default "bytes-per-reducer" threshold, etc.
Reference:
http://fr.slideshare.net/oom65/orc-andvectorizationhadoopsummit
https://cwiki.apache.org/confluence/display/Hive/LanguageManual+ORC
https://streever.atlassian.net/wiki/display/HADOOP/Optimizing+ORC+Files+for+Query+Performance
http://thinkbig.teradata.com/hadoop-performance-tuning-orc-snappy-heres-youre-missing/
One last thing: with 15 nodes for running queries and a replication factor of 3, each HDFS block is available "locally" on 3 the nodes (20%) and "remotely" in the rest (80%). A higher replication factor may reduce I/O and network bottlenecks -- at the cost of disk space, of course.
One of my Clients has a reservation based system. Similar to air lines. Running on MS SQL 2005.
The way the previous company has designed it is to create an allocation as a set of rows.
Simple Example Being:
AllocationId | SeatNumber | IsSold
1234 | A01 | 0
1234 | A02 | 0
In the process of selling a seat the system will establish an update lock on the table.
We have a problem at the moment where the locking process is running slow and we are looking at ways to speed it up.
The table is already efficiently index, so we are looking at a hardware solution to speed up the process. The table is about 5 mil active rows and sits on a RAID 50 SAS array.
I am assuming hard disk seek time is going to be the limiting factor in speeding up update locks when you have 5mil rows and are updating 2-5 rows at a time (I could be wrong).
I've herd about people using index partition over several disk arrays, has anyone had similar experiences with trying to speed up locking? can anyone give me some advise onto a possible solution on what hardware might be able to be upgraded or what technology we can take advantage of in order to speed up the update locks (without moving to a cluster)?
One last try…
It is clear that there are too many locks hold for too long.
Once the system starts slowing down
due to too many locks there is no
point in starting more transactions.
Therefore you should benchmark the system to find out the optimal number of currant transaction, then use some queue system (or otherwise) to limit the number of currant transaction. Sql Server may have some setting (number of active connections etc) to help, otherwise you will have to write this in your application code.
Oracle is good at allowing reads to bypass writes, however SqlServer is not as standared...
Therefore I would split the stored proc to use two transactions, the first transaction should just:
be a SNAPSHOT (or READ UNCOMMITTED) transaction
find the “Id” of the rows for the seats you wish to sell.
You should then commit (or abort) this transaction,
and use a 2nd (hopefully very short) transaction that
Most likcly is READ COMMITTED, (or maybe SERIALIZABLE)
Selects each row for update (use a locking hint)
Check it has not been sold in the mean time (abort and start again if it has)
Set the “IsSold” flag on the row
(You may be able to the above in a single update statement using “in”, and then check that the expected number of rows were updated)
Sorry sometimes you do need to understant what each time of transaction does and how locking works in detail.
If the table is smaller, then the
update is shorter and the locks are
hold for less time.
Therefore consider splitting the table:
so you have a table that JUST contains “AllocationId” and “IsSold”.
This table could be stored as a single btree (index organized table on AllocationId)
As all the other indexes will be on the table that contrains the details of the seat, no indexes should be locked by the update.
I don't think you'd getting anything out of table partitioning -- the only improvement you'd get would be in fewer disk reads from a smaller (shorter) index trees (each read will hit each level of the index at least once, so the fewer levels the quicker the read.) However, I've got a table with a 4M+ row partition, indexed on 4 columns, net 10 byte key length. It fits in three index levels, with the topmost level 42.6% full. Assuming you had something similar, it seems reasonable that partitioning might only remove one level from the tree, and I doubt that's much of an improvement.
Some off the-cuff hardward ideas:
Raid 5 (and 50) can be slower on writes, because of the parity calculation. Not an issue (or so I'm told) if the disk I/O cache is large enough to handle the workload, but if that's flooded you might want to look at raid 10.
Partition the table across multiple drive arrays. Take two (or more) Raid arrays, distribute the table across the volumes[files/file groups, with or without table partitioning or partitioned views], and you've got twice the disk I/O speed, depending on where the data lies relative to the queries retrieving it. (If everythings on array #1 and array #2 is idle, you've gained nothing.)
Worst case, there's probably leading edge or bleeding edge technology out there that will blow your socks off. If it's critical to your business and you've got the budget, might be worth some serious research.
How long is the update lock hold for?
Why is the lock on the “table” not just the “rows” being sold?
If the lock is hold for more then a
faction of a second that is likely to
be your problem. SqlServer does not
like you holding locks while users
fill in web forms etc.
With SqlServer, you have to implement a “shopping cart” yourself, by temporary reserving the seat until the user pays for it. E.g add a “IsReserved” and “ReservedAt” colunn, then any seats that has been reserved for more then n minutes should be automatically unreserved.
This is a hard problem, as a shopper does not expect a seat that is in stock to be sold to someone else where he is checking out. However you don’t know if the shopper will ever complete the checkout. So how do you show it on a UI. Think about having a look at what other booking websites do then copy one that your users already know how to use.
(Oracle can sometimes cope with lock being kept for a long time, but even Oracle is a lot faster and happier if you keep your locking short.)
I would first try to figure out why the you are locking the table rather than just a row.
One thing to check out is the Execution plan of the Update statement to see what Indexes it causes to be updated and then make sure that row_level_lock and page_level_lock are enabled on those indexes.
You can do so with the following statement.
Select allow_row_locks, allow_page_locks from sys.indexes where name = 'IndexNameHere'
Here are a few ideas:
Make sure your data and logs are on separate spindles, to maximize write performance.
Configure your drives to only use the first 30% or so for data, and have the remainder be for backups (minimize seek / random access times).
Use RAID 10 for the log volume; add more spindles as needed for performance (write performance is driven by the speed of the log)
Make sure your server has enough RAM. Ideally, everything needed for a transaction should be in memory before the transaction starts, to minimize lock times (consider pre-caching). There are a bunch of performance counters you can check for this.
Partitioning may help, but it depends a lot on the details of your app and data...
I'm assuming that the T-SQL, indexes, transaction size, etc, have already been optimized.
In case it helps, I talk about this subject in detail in my book (including SSDs, disk array optimization, etc) -- Ultra-Fast ASP.NET.