Problem
I need to establish write consistency for a sequence of queries using updateMany, against a DocumentDB cluster with only a single primary instance. I am not sure which approach to use, between Transactions, ordered BulkWrites, or simply setting a Majority write concern for each updateMany query.
Environment
AWS DocumentDB cluster, which maps to MongoDB v4.0, via pymongo 3.12.0 .
Note: the cluster has a single primary instance, and no other instances. In practice, AWS will have us connect to the cluster in replica set mode. I am not sure whether this means we need to still think about this problem in terms of replica sets.
Description
I have a sequence of documents D , each of which is an array of records. Each record is of the form {field: MyField, from_id: A, to_id: B}.
To process a record, I need to look in my DB for all fields MyField that have value A, and then set that value to B. The actual query I use to do this is updateMany. The code looks something like:
for doc in Documents:
for record in doc:
doWriteUpdate(record)
def doWriteUpdate(record):
query = ... # format the query based on record's information
db.updateMany(query)
I need the update operations to happen such that the writes have actually been applied, and are visible, before the next doWriteUpdate query runs.
This is because I expect to encounter a situation where I can have a record {field: MyField, from_id: A, to_id: B}, and then a subsequent record (whether in the same document, or a following document) {field: MyField, from_id: B, to_id: C}. Being able to properly apply the latter record operation, depends on the former record operation having been committed to the database.
Possible Approaches
Transactions
I have tried wrapping my updateMany operation in a Transaction. If this had worked, I would have called it a day; but I exceed the size allowed: Total size of all transaction operations must be less than 33554432. Without rewriting the queries, this cannot be worked around, because the updateMany has several layers of array-filtering, and digs through a lot of documents. I am not even sure if transactions are appropriate in this case, because I am not using any replica sets, and they seem to be intended for ACID with regard to replication.
Ordered Bulk Writes
BulkWrite.updateMany would appear to guarantee execution order of a sequence of writes. So, one approach could be, to generate the update query strings for each record r in a document D, and then send those through (preserving order) as a BulkWrite. While this would seem to "preserve order" of execution, I don't know if a) the preservation of execution order, also guarantees write consistency (everything executed serially is applied serially), and, more important, b) whether the following BulkWrites, for the other documents, will interleave with this one.
WriteConcern
Pymongo states that writes will block given a desired WriteConcern. My session is single-threaded, so this should give the desired behavior. However, MongoDB says
For multi-document transactions, you set the write concern at the transaction level, not at the individual operation level. Do not explicitly set the write concern for individual write operations in a transaction.
I am not clear on whether this pertains to "transactions" as in the general sense, or MongoDB Transactions set up through session objects. If it means the latter, then it shouldn't apply to my use case. If the former, then I don't know what other approach to use.
The proper write concern is majority, and with a read concern that uses the linearizable
Real Time Order Combined with "majority" write concern, "linearizable"
read concern enables multiple threads to perform reads and writes on a
single document as if a single thread performed these operations in
real time; that is, the corresponding schedule for these reads and
writes is considered linearizable.
Related
When we run a Mongo find() query without any sort order specified, what does the database internally use to sort the results?
According to the documentation on the mongo website:
When executing a find() with no parameters, the database returns
objects in forward natural order.
For standard tables, natural order is not particularly useful because,
although the order is often close to insertion order, it is not
guaranteed to be. However, for Capped Collections, natural order is
guaranteed to be the insertion order. This can be very useful.
However for standard collections (non capped collections), what field is used to sort the results?
Is it the _id field or something else?
Edit:
Basically, I guess what I am trying to get at is that if I execute the following search query:
db.collection.find({"x":y}).skip(10000).limit(1000);
At two different points in time: t1 and t2, will I get different result sets:
When there have been no additional writes between t1 & t2?
When there have been new writes between t1 & t2?
There are new indexes that have been added between t1 & t2?
I have run some tests on a temp database and the results I have gotten are the same (Yes) for all the 3 cases - but I wanted to be sure and I am certain that my test cases weren't very thorough.
What is the default sort order when none is specified?
The default internal sort order (or natural order) is an undefined implementation detail. Maintaining order is extra overhead for storage engines and MongoDB's API does not mandate predictability outside of an explicit sort() or the special case of fixed-sized capped collections which have associated usage restrictions. For typical workloads it is desirable for the storage engine to try to reuse available preallocated space and make decisions about how to most efficiently store data on disk and in memory.
Without any query criteria, results will be returned by the storage engine in natural order (aka in the order they are found). Result order may coincide with insertion order but this behaviour is not guaranteed and cannot be relied on (aside from capped collections).
Some examples that may affect storage (natural) order:
WiredTiger uses a different representation of documents on disk versus the in-memory cache, so natural ordering may change based on internal data structures.
The original MMAPv1 storage engine (removed in MongoDB 4.2) allocates record space for documents based on padding rules. If a document outgrows the currently allocated record space, the document location (and natural ordering) will be affected. New documents can also be inserted in storage marked available for reuse due to deleted or moved documents.
Replication uses an idempotent oplog format to apply write operations consistently across replica set members. Each replica set member maintains local data files that can vary in natural order, but will have the same data outcome when oplog updates are applied.
What if an index is used?
If an index is used, documents will be returned in the order they are found (which does necessarily match insertion order or I/O order). If more than one index is used then the order depends internally on which index first identified the document during the de-duplication process.
If you want a predictable sort order you must include an explicit sort() with your query and have unique values for your sort key.
How do capped collections maintain insertion order?
The implementation exception noted for natural order in capped collections is enforced by their special usage restrictions: documents are stored in insertion order but existing document size cannot be increased and documents cannot be explicitly deleted. Ordering is part of the capped collection design that ensures the oldest documents "age out" first.
It is returned in the stored order (order in the file), but it is not guaranteed to be that they are in the inserted order. They are not sorted by the _id field. Sometimes it can be look like it is sorted by the insertion order but it can change in another request. It is not reliable.
Imagine theres a document containing a single field: {availableSpots: 100}
and there are millions of users, racing to get a spot by sending a request to an API server.
each time a request comes, the server reads the document and if the availableSpot is > 0, it then decrements it by 1 and creates a booking in another collection.
Now i read that mongodb locks the document whenever an update operation is performed.
What will happen if theres a million concurrent requests? will it take a long time because the same document keeps getting locked? Also, the server reads the value of document before it tries to update the document, and by the time it acquires the lock, the spot may not be available anymore.
It is also possible that the threads are getting "availableSpot > 0" is true at the same instant in time, but in reality the availableSpot may not be enough for all the requests. How to deal with this?
The most important thing here is atomicity and concurrency.
1. Atomicity
Your operation to update (decrement by one) if availableSpots > 0 :
db.collection.updateOne({"availableSpots" :{$gt : 0}}, { $inc: { availableSpots: -1 })
is atomic.
$inc is an atomic operation within a single document.
Refer : https://docs.mongodb.com/manual/reference/operator/update/inc/
2. Concurrency
Since MongoDB has document-level concurrency control for write operations. Each update will take a lock on the document.
Now your questions:
What will happen if theres a million concurrent requests?
Yes each update will be performed one by one (due to locking) hence will slow down.
the server reads the value of document before it tries to update the
document, and by the time it acquires the lock, the spot may not be
available anymore.
Since the operation is atomic, this will not happen. It will work as you want, only 100 updates will be executed with number of affected rows greater than 0 or equal to 1.
MongoDB uses Wired Tiger as a default storage engine starting version 3.2.
Wired Tiger provides document level concurrency:
From docs:
WiredTiger uses document-level concurrency control for write
operations. As a result, multiple clients can modify different
documents of a collection at the same time.
For most read and write operations, WiredTiger uses optimistic
concurrency control. WiredTiger uses only intent locks at the global,
database and collection levels. When the storage engine detects
conflicts between two operations, one will incur a write conflict
causing MongoDB to transparently retry that operation.
When multiple clients are trying to update a value in a document, only that document will be locked, but not the entire collections.
My understanding is that you are concerned about the performance of many concurrent ACID-compliant transactions against two separate collections:
a collection (let us call it spots) with one document {availableSpots: 999..}
another collection (let us call it bookings) with multiple documents, one per booking.
Now i read that mongodb locks the document whenever an update operation is performed.
It is also possible that the threads are getting "availableSpot > 0"
is true at the same instant in time, but in reality the availableSpot
may not be enough for all the requests. How to deal with this?
With version 4.0, MongoDB provides the ability to perform multi-document transactions against replica sets. (The forthcoming MongoDB 4.2 will extend this multi-document ACID transaction capability to sharded clusters.)
This means that no write operations within a multi-document transaction (such as updates to both the spots and bookings collections, per your proposed approach) are visible outside the transaction until the transaction commits.
Nevertheless, as noted in the MongoDB documentation on transactions a denormalized approach will usually provide better performance than multi-document transactions:
In most cases, multi-document transaction incurs a greater performance
cost over single document writes, and the availability of
multi-document transaction should not be a replacement for effective
schema design. For many scenarios, the denormalized data model
(embedded documents and arrays) will continue to be optimal for your
data and use cases. That is, for many scenarios, modeling your data
appropriately will minimize the need for multi-document transactions.
In MongoDB, an operation on a single document is atomic. Because you can use embedded documents and arrays to capture relationships between data in a single document structure instead of normalizing across multiple documents and collections, this single-document atomicity obviates the need for multi-document transactions for many practical use cases.
But do bear in mind that your use case, if implemented within one collection as a single denormalized document containing one availableSpots sub-document and many thousands of bookings sub-documents, may not be feasible as the maximum document size is 16MB.
So, in conclusion, a denormalized approach to write atomicity will usually perform better than a multi-document approach, but is constrained by the maximum document size of 16MB.
You can try using findAndModify() option while trying to update the document. Each time you will need to cherry pick whichever field you want to update in that particular document. Also, since mongo db replicates data to Primary and secondary nodes, you may also want to adjust your WriteConcern values as well. You can read more about this in official documentation. I have something similar coded that handles similar kind of concurrency issues in mongoDB using spring mongoTemplate. Let me know if you want any reference related to java with that.
When we run a Mongo find() query without any sort order specified, what does the database internally use to sort the results?
According to the documentation on the mongo website:
When executing a find() with no parameters, the database returns
objects in forward natural order.
For standard tables, natural order is not particularly useful because,
although the order is often close to insertion order, it is not
guaranteed to be. However, for Capped Collections, natural order is
guaranteed to be the insertion order. This can be very useful.
However for standard collections (non capped collections), what field is used to sort the results?
Is it the _id field or something else?
Edit:
Basically, I guess what I am trying to get at is that if I execute the following search query:
db.collection.find({"x":y}).skip(10000).limit(1000);
At two different points in time: t1 and t2, will I get different result sets:
When there have been no additional writes between t1 & t2?
When there have been new writes between t1 & t2?
There are new indexes that have been added between t1 & t2?
I have run some tests on a temp database and the results I have gotten are the same (Yes) for all the 3 cases - but I wanted to be sure and I am certain that my test cases weren't very thorough.
What is the default sort order when none is specified?
The default internal sort order (or natural order) is an undefined implementation detail. Maintaining order is extra overhead for storage engines and MongoDB's API does not mandate predictability outside of an explicit sort() or the special case of fixed-sized capped collections which have associated usage restrictions. For typical workloads it is desirable for the storage engine to try to reuse available preallocated space and make decisions about how to most efficiently store data on disk and in memory.
Without any query criteria, results will be returned by the storage engine in natural order (aka in the order they are found). Result order may coincide with insertion order but this behaviour is not guaranteed and cannot be relied on (aside from capped collections).
Some examples that may affect storage (natural) order:
WiredTiger uses a different representation of documents on disk versus the in-memory cache, so natural ordering may change based on internal data structures.
The original MMAPv1 storage engine (removed in MongoDB 4.2) allocates record space for documents based on padding rules. If a document outgrows the currently allocated record space, the document location (and natural ordering) will be affected. New documents can also be inserted in storage marked available for reuse due to deleted or moved documents.
Replication uses an idempotent oplog format to apply write operations consistently across replica set members. Each replica set member maintains local data files that can vary in natural order, but will have the same data outcome when oplog updates are applied.
What if an index is used?
If an index is used, documents will be returned in the order they are found (which does necessarily match insertion order or I/O order). If more than one index is used then the order depends internally on which index first identified the document during the de-duplication process.
If you want a predictable sort order you must include an explicit sort() with your query and have unique values for your sort key.
How do capped collections maintain insertion order?
The implementation exception noted for natural order in capped collections is enforced by their special usage restrictions: documents are stored in insertion order but existing document size cannot be increased and documents cannot be explicitly deleted. Ordering is part of the capped collection design that ensures the oldest documents "age out" first.
It is returned in the stored order (order in the file), but it is not guaranteed to be that they are in the inserted order. They are not sorted by the _id field. Sometimes it can be look like it is sorted by the insertion order but it can change in another request. It is not reliable.
This is more of 'inner workings' undestanding question:
How do noSQL databases that do not support *A*CID (meaning that they cannot update/insert and then rollback data for more than one object in a single transaction) -- update the secondary indexes ?
My understanding is -- that in order to keep the secondary index in sync (other wise it will become stale for reads) -- this has to happen withing the same transaction.
furthermore, if it is possible for index to reside on a different host than the data -- then a distributed lock needs to be present and/or two-phase commit for such an update to work atomically.
But if these databases do not support the multi-object transactions (which means they do not do two-phase commit on data across multiple host) , what method do they use to guarantee that secondary indices that reside in B-trees structures separate from the data are not stale ?
This is a great question.
RethinkDB always stores secondary indexes on the same host as the primary index/data for the table. Even in case of joins, RethinkDB brings the query to the data, so the secondary indexes, primary indexes, and data always reside on the same node. As a result, there is no need for distributed locking protocols such as two phase commit.
RethinkDB does support a limited set of transactional functionality -- single document transactions. Changes to a single document are recorded atomically. Relevant secondary index changes are also recorded as part of that transaction, so either the entire change is recorded, or nothing is recorded at all.
It would be easy to extend the limited transactional functionality to support multiple documents in a single shard, but it would be hard to do it across shards (for the distributed locking reasons you brought up), so we decided not to implement transactions for multiple documents yet.
Hope this helps.
This is a MongoDB answer.
I am not quite sure what your logic here is. Updating a secondary index has nothing to do with being able to rollback multi statement transactions such as a multiple update.
MongoDB has transcactions per a single document, and that is what matters for updating indexes. These operations can be reversed using the journal if the need arises.
this has to happen withing the same transaction.
Yes, much like a RDBMS would. The more indexes you apply the slower your writes will be, and it seems to me you know why.
As the write occurs MongoDB will update all indexes which apply to that collection with the fields that apply to specific indexes.
furthermore, if it is possible for index to reside on a different host than the data
I am unsure if MongoDB allows that, I believe there is a JIRA for it; however, I cannot find that JIRA currently.
then a distributed lock needs to be present and/or two-phase commit for such an update to work atomically.
Most likely. Allowing this feature would be...well, let's just say creating a hairball.
Even in a sharded setup the index of each range resides on the shard itself, not on the config servers.
But if these databases do not support the multi-object transactions (which means they do not do two-phase commit on data across multiple host)
That is not what a two phase commit means. I believe you need to brush up on what a two phase commit is: http://docs.mongodb.org/manual/tutorial/perform-two-phase-commits/
I suppose if you are talking about a transaction covering more than one shard then, hmm ok.
what method do they use to guarantee that secondary indices that reside in B-trees structures separate from the data are not stale ?
Agan I am unsure why a multi document transaction would effect whether an index would be stale or not, your not grouping across documents. The exception to that is a unique index but that works on single document updates as well; note that its uniqueness gets kinda hairy in sharded setups and cannot be guaranteed.
In an index you are creating, normally, one entry per document prefix key, uless it is a multikey index on the docment then you can make more than one index, however, either way index updating is done per single object, not by multi document transactions and I am unsure what you logic here is aas such this is the answer I have placed.
RethinkDB always stores secondary index data on the same machine as the data it's indexing. This allows it to be updated within the same transaction. Rethink promises to be ACIDy with single document operations and considers the indexing of a document to be part of the document itself.
When we run a Mongo find() query without any sort order specified, what does the database internally use to sort the results?
According to the documentation on the mongo website:
When executing a find() with no parameters, the database returns
objects in forward natural order.
For standard tables, natural order is not particularly useful because,
although the order is often close to insertion order, it is not
guaranteed to be. However, for Capped Collections, natural order is
guaranteed to be the insertion order. This can be very useful.
However for standard collections (non capped collections), what field is used to sort the results?
Is it the _id field or something else?
Edit:
Basically, I guess what I am trying to get at is that if I execute the following search query:
db.collection.find({"x":y}).skip(10000).limit(1000);
At two different points in time: t1 and t2, will I get different result sets:
When there have been no additional writes between t1 & t2?
When there have been new writes between t1 & t2?
There are new indexes that have been added between t1 & t2?
I have run some tests on a temp database and the results I have gotten are the same (Yes) for all the 3 cases - but I wanted to be sure and I am certain that my test cases weren't very thorough.
What is the default sort order when none is specified?
The default internal sort order (or natural order) is an undefined implementation detail. Maintaining order is extra overhead for storage engines and MongoDB's API does not mandate predictability outside of an explicit sort() or the special case of fixed-sized capped collections which have associated usage restrictions. For typical workloads it is desirable for the storage engine to try to reuse available preallocated space and make decisions about how to most efficiently store data on disk and in memory.
Without any query criteria, results will be returned by the storage engine in natural order (aka in the order they are found). Result order may coincide with insertion order but this behaviour is not guaranteed and cannot be relied on (aside from capped collections).
Some examples that may affect storage (natural) order:
WiredTiger uses a different representation of documents on disk versus the in-memory cache, so natural ordering may change based on internal data structures.
The original MMAPv1 storage engine (removed in MongoDB 4.2) allocates record space for documents based on padding rules. If a document outgrows the currently allocated record space, the document location (and natural ordering) will be affected. New documents can also be inserted in storage marked available for reuse due to deleted or moved documents.
Replication uses an idempotent oplog format to apply write operations consistently across replica set members. Each replica set member maintains local data files that can vary in natural order, but will have the same data outcome when oplog updates are applied.
What if an index is used?
If an index is used, documents will be returned in the order they are found (which does necessarily match insertion order or I/O order). If more than one index is used then the order depends internally on which index first identified the document during the de-duplication process.
If you want a predictable sort order you must include an explicit sort() with your query and have unique values for your sort key.
How do capped collections maintain insertion order?
The implementation exception noted for natural order in capped collections is enforced by their special usage restrictions: documents are stored in insertion order but existing document size cannot be increased and documents cannot be explicitly deleted. Ordering is part of the capped collection design that ensures the oldest documents "age out" first.
It is returned in the stored order (order in the file), but it is not guaranteed to be that they are in the inserted order. They are not sorted by the _id field. Sometimes it can be look like it is sorted by the insertion order but it can change in another request. It is not reliable.