We use tarantool database (sharded using vshard) in production. We started directly with 4 shards. Now we want to increase it to 6 without downtime. But, after adding two more shards, rebalancer kicks in and it doesn't allow reads/writes to happen. Is there any way, that rebalancing can happen supporting all kinds of operations? We can afford to increase the operation time. But it should be a success. What is the best practice to add a shard to tarantool with the minimum inconvenience caused in the product front?
Currently, the Only solution we can think of is to go into maintenance mode and have the rebalance to finish with minimum time possible!!!
You can not write to a bucket that is being transferred right now, but you cant write to other buckets (so it's not like the whole shard is locked up).
Moreover, you can mitigate the effect by
- making buckets smaller (increase bucket_count)
- making rebalancing slower so that that less buckets are transferred simultaneoulsy (rebalancer config).
Suppose, you have 16384 buckets and your dataset is 75GB. It means that average bucket size is around 5 Mb. If you decrease rebalancer_max_receiving parameter to 10, you'll have only 10 buckets (50Mb) being transferred simultaneously (which makes him locked for writes).
This way, rebalancing will be pretty slow, BUT, given that your clients can perform retries and your network between shards is fast enough, the 'write-lock' effect should got unnoticed at all.
Related
When I am adding 200 documents to ElasticSearch via one bulk request - it's super fast.
But I am wondering if is there a chance to speed up the process with concurrent executions: 20 concurrent executions with 10 documents each.
I know it's not efficient, but maybe there is a chance to speed up the process with concurrent executions?
Lower concurrency is preferable for bulk document inserts. Some concurrency is helpful in some circumstances — It Depends™ and I'll get into it — but is not a major or automatic win.
There's a lot that can be tuned when it comes to performance of writes to Elasticsearch. One really quick win that you should check: are you using HTTP keep-alive for your connections? That's going to save a lot of the TCP and TLS overhead of setting up each connection. Just that change can make a big performance boost, and also uncover some meaningful architectural considerations for your indexing pipeline.
So check that out and see how it goes. From there, we should go to the bottom, and work our way up.
The index on disk is Lucene. Lucene is a segmented index. The index part is a core reason why you're using Elasticsearch in the first place: a dictionary of sorted terms can be searched in O(log N) time. That's super fast and scalable. The segment part is because inserting into an index is not particularly fast — depending on your implementation, it costs O(log N) or O(N log N) to maintain the sorting.
So Lucene's trick is to buffer those updates and append a new segment; essentially a collection of mini-indices. Searching some relatively small number of segments is still much faster than taking all the time to maintain a sorted index with every update. Over time Lucene takes care of merging these segments to keep them within some sensible size range, expunging deleted and overwritten docs in the process.
In Elasticsearch, every shard is a distinct Lucene index. If you have an index with a single shard, then there is very little benefit to having more than a single concurrent stream of bulk updates. There may be some benefit to concurrency on the application side, depending on the amount of time it takes for your indexing pipeline to collect and assemble each batch of documents. But on the Elasticsearch side, it's all just one set of buffers getting written out to one segment after another.
Sharding makes this a little more interesting.
One of Elasticsearch's strengths is the ability to partition the data of an index across multiple shards. This helps with availability, and it helps workloads scale beyond the resources of a single server.
Alas it's not quite so simple as to say that the concurrency should be equal, or proportional, to the number of primary shards that an index has. Although, as a rough heuristic, that's not a terrible one.
You see, internally, the first Elasticsearch node to handle the request is going to turn that Bulk request into a sequence of individual document update actions. Each document update is sent to the appropriate node that is hosting the shard that this document belongs to. Responses are collected by the bulk action so that it can send a summary of the bulk operation in its response to the client.
So at this point, depending on the document-shard routing, some shards may be busier than others during the course of processing an incoming bulk request. Is that likely to matter? My intuition says not really. It's possible, but it would be unusual.
In most tests and analysis I've seen, and in my experience over ~ten years with Lucene, the slow part of indexing is the transformation of the documents' values into the inverted index format. Parsing the text, analyzing it into terms, and so on, can be very complex and costly. So long as a bulk request has sufficient documents that are sufficiently well distributed across shards, the concurrency is not as meaningful as saturating the work done at the shard and segment level.
When tuning bulk requests, my advice is something like this.
Use HTTP keep-alive. This is not optional. (You are using TLS, right?)
Choose a batch size where each request is taking a modest amount of time. Somewhere around 1 second, probably not more than 10 seconds.
If you can get fancy, measure how much time each bulk request took, and dynamically grow and shrink your batch.
A durable queue unlocks a lot of capabilities. If can fetch and assemble documents and insert them into, say, Kafka, then that process can be run in parallel to saturate the database and parallelize any denormalization or preparation of documents. A different process then pulls from the queue and sends requests to the server, and with some light coordination you can test and tune different concurrencies at different stages. A queue also lets you pause your updates for various migrations and maintenance tasks when it helps to put the cluster into read-only mode for a time.
I've avoided replication throughout this answer because there's only one reason where I'd ever recommend tweaking replication. And that is when you are bulk creating an index that is not serving any production traffic. In that case, it can help save some resources through your server fleet to turn off all replication to the index, and enable replication after the index is essentially done being loaded with data.
To close, what if you crank up the concurrency anyway? What's the risk? Some workloads don't control the concurrency and there isn't the time or resources to put a queue in front of the search engine. In that case, Elasticsearch can avoid a fairly substantial amount of concurrency. It has fairly generous thread pools for handling concurrent document updates. If those thread pools are saturated, it will reject responses with a HTTP 429 error message and a clear message about queue depths being exceeded. Those can impact stability of the cluster, depending on available resources, and number of shards in the index. But those are all pretty noticeable issues.
Bottom line: no, 20 concurrent bulks with 10 documents each will probably not speed up performance relative to 1 bulk with 200 documents. If your bulk operations are fast, you should increase their size until they run for a second or two, or are problematic. Use keep-alive. If there is other app-side overhead, increase your concurrency to 2x or 3x and measure empirically. If indexing is mission critical, use a fast, durable queue.
There is no straight answer to this as it depends on lots of factors. Above the optimal bulk request size, performance no longer improves and may even drop off. The optimal size, however, is not a fixed number.
It depends entirely on your hardware, your document size and complexity, and your indexing and search load.
Try indexing typical documents in batches of increasing size. When performance starts to drop off, your batch size is too big.
Since you are doing it in batches of 200, the chances are high that it should be most optimal way to index. But again it will depend on the factors mentioned above.
I am using crate 1.0.2 which internally uses elasticsearch. So my question is applicable for both. For certain queries I get circuit break exception.
CircuitBreakingException: [parent] Data too large, data for [collect: 0] would be larger than limit of [11946544332/11.1gb]
These queries are mainly group by on multiple columns. I have billions of documents indexed and have 16 GB of RAM allocated as crate heap size. I have multiple such nodes connected together in a cluster. Will adding more nodes in the cluster help in getting rid of this error and will my same queries run fine ? Or is it that I must increase heap to 30 GB ? My worry is when I increase it to 30GB and as I add more data, someday that query will again hit the circuit breaker. So I wanted to solve it by scaling horizontally i.e. adding more nodes. Will that be wiser decision ?
Short answer: Usually horizontal scaling helps.
Your error seems to be caused by group by queries.
The GROUP BY operations are executed in a distributed fashion. So more nodes
will generally split the load and therefore also the memory usage. (Make sure
there are enough shards so that they're spread among all nodes)
There is a catch though: Eventually the data needs to be merged together on the
node you sent the initial query to. This is generally fine because the data
arrives pre-aggregated, but If the cardinality is too high (Ex. GROUP BY on a
primary key), the whole data set has to fit into memory on this coordinator
node.
If your nodes have enough memory to go up to 30 GB (with still having enough to
spare for the file system cache), I'd personally tend to increase the HEAP size
first, before adding new nodes.
Update:
Recent versions (2.1.X) also contain some fixes regarding the circuit-breaker behaviur. So if it's possible to update that'd be recommended as well.
Update2:
Note that there are different cases in which a circuit breaker can trip. In
your case it's caused by a GROUP BY using up quite a lot of memory. But it can
also be tripped if a single request is too large. For example if the bulk size
is too large. In such a case more nodes wouldn't help. You'd have to reduce the
bulk size.
My question is mostly based on the following article:
https://qbox.io/blog/optimizing-elasticsearch-how-many-shards-per-index
The article advises against having multiple shards per node for two reasons:
Each shard is essentially a Lucene index, it consumes file handles, memory, and CPU resources
Each search request will touch a copy of every shard in the index. Contention arises and performance decreases when the shards are competing for the same hardware resources
The article advocates the use of rolling indices for indices that see many writes and fewer reads.
Questions:
Do the problems of resource consumption by Lucene indices arise if the old indices are left open?
Do the problems of contention arise when searching over a large time range involving many indices and hence many shards?
How does searching many small indices compare to searching one large one?
I should mention that in our particular case, there is only one ES node though of course generally applicable answers will be more useful to SO readers.
It's very difficult to spit out general best practices and guidelines when it comes to cluster sizing as it depends on so many factors. If you ask five ES experts, you'll get ten different answers.
After several years of tinkering and fiddling around ES, I've found out that what works best for me is always to start small (one node, how many indices your app needs and one shard per index), load a representative data set (ideally your full data set) and load test to death. Your load testing scenarii should represent the real maximum load you're experiencing (or expecting) in your production environment during peak hours.
Increase the capacity of your cluster (add shard, add nodes, tune knobs, etc) until your load test pass and make sure to increase your capacity by a few more percent in order to allow for future growth. You don't want your production to be fine now, you want it to be fine in a year from now. Of course, it will depend on how fast your data will grow and it's very unlikely that you can predict with 100% certainty what will happen in a year from now. For that reason, as soon as my load test pass, if I expect a large exponential data growth, I usually increase the capacity by 50% more percent, knowing that I will have to revisit my cluster topology within a few month or a year.
So to answer your questions:
Yes, if old indices are left open, they will consume resources.
Yes, the more indices you search, the more resources you will need in order to go through every shard of every index. Be careful with aliases spanning many, many rolling indices (especially on a single node)
This is too broad to answer, as it again depends on the amount of data we're talking about and on what kind of query you're sending, whether it uses aggregation, sorting and/or scripting, etc
Do the problems of resource consumption by Lucene indices arise if the old indices are left open?
Yes.
Do the problems of contention arise when searching over a large time range involving many indices and hence many shards?
Yes.
How does searching many small indices compare to searching one large one?
When ES searches an index it will pick up one copy of each shard (be it replica or primary) and asks that copy to run the query on its own set of data. Searching a shard will use one thread from the search threadpool the node has (the threadpool is per node). One thread basically means one CPU core. If your node has 8 cores then at any given time the node can search concurrently 8 shards.
Imagine you have 100 shards on that node and your query will want to search all of them. ES will initiate the search and all 100 shards will compete for the 8 cores so some shards will have to wait some amount of time (microseconds, milliseconds etc) to get their share of those 8 cores. Having many shards means less documents on each and, thus, potentially a faster response time from each. But then the node that initiated the request needs to gather all the shards' responses and aggregate the final result. So, the response will be ready when the slowest shard finally responds with its set of results.
On the other hand, if you have a big index with very few shards, there is not so much contention for those CPU cores. But the shards having a lot of work to do individually, it can take more time to return back the individual result.
When choosing the number of shards many aspects need to be considered. But, for some rough guidelines yes, 30GB per shard is a good limit. But this won't work for everyone and for every use case and the article fails to mention that. If, for example, your index is using parent/child relationships those 30GB per shard might be too much and the response time of a single shard can be too slow.
You took this out of the context: "The article advises against having multiple shards per node". No, the article advises one to think about the aspects of structuring the indices shards before hand. One important step here is the testing one. Please, test your data before deciding how many shards you need.
You mentioned in the post "rolling indices", and I assume time-based indices. In this case, one question is about the retention period (for how long you need the data). Based on the answer to this question you can determine how many indices you'll have. Knowing how many indices you'll have gives you the total number of shards you'll have.
Also, with rolling indices, you need to take care of deleting the expired indices. Have a look at Curator for this.
I understand that there are two dominant patterns for keeping a rolling window of data inside elasticsearch:
creating daily indices, as suggested by logstash, and dropping old indices, and therefore all the records they contain, when they fall out of the window
using elasticsearch's TTL feature and a single index, having elasticsearch automatically remove old records individually as they fall out of the window
Instinctively I go with 2, as:
I don't have to write a cron job
a single big index is easier to communicate to my colleagues and for them to query (I think?)
any nightmare stream dynamics, that cause old log events to show up, don't lead to the creation of new indices and the old events only hang around for the 60s period that elasticsearch uses to do ttl cleanup.
But my gut tells me that dropping an index at a time is probably a lot less computationally intensive, though tbh I've no idea how much less intensive, nor how costly the ttl is.
For context, my inbound streams will rarely peak above 4K messages per second (mps) and are much more likely to hang around 1-2K mps.
Does anyone have any experience with comparing these two approaches? As you can probably tell I'm new to this world! Would appreciate any help, including even help with what the correct approach is to thinking about this sort of thing.
Cheers!
Short answer is, go with option 1 and simply delete indexes that are no longer needed.
Long answer is it somewhat depends on the volume of documents that you're adding to the index and your sharding and replication settings. If your index throughput is fairly low, TTLs can be performant but as you start to write more docs to Elasticsearch (or if you a high replication factor) you'll run into two issues.
Deleting documents with a TTL requires that Elasticsearch runs a periodic service (IndicesTTLService) to find documents that are expired across all shards and issue deletes for all those docs. Searching a large index can be a pretty taxing operation (especially if you're heavily sharded), but worse are the deletes.
Deletes are not performed instantly within Elasticsearch (Lucene, really) and instead documents are "marked for deletion". A segment merge is required to expunge the deleted documents and reclaim disk space. If you have large number of deletes in the index, it'll put much much more pressure on your segment merge operations to the point where it will severely affect other thread pools.
We originally went the TTL route and had an ES cluster that was completely unusable and began rejecting search and indexing requests due to greedy merge threads.
You can experiment with "what document throughput is too much?" but judging from your use case, I'd recommend saving some time and just going with the index deletion route which is much more performant.
I would go with option 1 - i.e. daily dropping of indices.
Daily Dropping Indices
pros:
This is the most efficient way of deleting data
If you need to restructure your index (e.g. apply a new mapping, increase number of shards) any changes are easily applied to the new index
Details of the current index (i.e. the name) is hidden from clients by using aliases
Time based searches can be directed to search only a specific small index
Index templates simplify the process of creating the daily index.
These benefits are also detailed in the Time-Based Data Guide, see also Retiring Data
cons:
Needs more work to set up (e.g. set up of cron jobs), but there is a plugin (curator) that can help with this.
If you perform updates on data then all versions of a document data will need to sit in the same index, i.e. multiple indexes won't work for you.
Use of TTL or Queries to delete data
pros:
Simple to understand and easily implemented
cons:
When you delete a document, it is only marked as deleted. It won’t be physically deleted until the segment containing it is merged away. This is very inefficient as the deleted data will consume disk space, CPU and memory.
I have a website for classified. For this I'm using elasticsearch, postgres and rails on a same ubuntu 14.04 dedicated server, with 256GB of RAM and 20 cores, 40 threads.
I have 10 indexes on elasticsearch, each have default numbers of shards (5). They have between 1000 and 400 000 classifieds depending on which index.
approximately 5000 requests per minute, 2/3 making an elasticsearch request.
according to htop, jvm is using around 500% of CPU
I try different options, I reduce number of shards per index, I also try to change JAVA_OPTS as followed
#JAVA_OPTS="$JAVA_OPTS -XX:+UseParNewGC"
#JAVA_OPTS="$JAVA_OPTS -XX:+UseConcMarkSweepGC"
#JAVA_OPTS="$JAVA_OPTS -XX:CMSInitiatingOccupancyFraction=75"
#JAVA_OPTS="$JAVA_OPTS -XX:+UseCMSInitiatingOccupancyOnly"
JAVA_OPTS="$JAVA_OPTS -XX:+UseG1GC"
but it doesn't seems to change anything.
so to questions :
when you change any setting on elasticsearch, and then restart, should the improvement (if any) be visible immediately or can it arrive a bit later thanks to cache or anything else ?
can any one help me to find good configuration for JVM / elasticsearch so it will not take that many resources
First, it's a horrible idea to run your web server, database and Elasticsearch server all on the same box. Each of these should be given it's own box, at least. In the case of Elasticsearch, it's actually recommended to have at least 3 servers, or nodes. That way you end up with a load balanced cluster that won't run into split-brain issues.
Further, sharding only makes sense in a cluster. If you only have one node, then all the shards reside on the same node. This causes two performance problems. First, you get the hit that sharding always adds. For every query, Elasticsearch must query each shard individually (each is a separate Lucene index). Then, it must combine and process the result from all the shards to produce the final result. That's a not insignificant amount of overhead. Second, because all the shards reside on the same node, you're I/O-locked. The shards have to be queried one at a time instead of all at once. Optimally, you should have one shard per node, however, since you can't create more shards without reindexing, it's common to have a few extra hanging around for future horizontal scaling. In that scenario, the cost of reindexing what could be 100's of gigs of data or more outweighs a little bit of performance bottleneck. However, if you've got 5 shards running one node, that's probably a large part of your performance problems right there.
Finally, and again, with Elasticsearch in particular, swapping is a huge no-no. Most of what makes Elasticsearch efficient is it's cache which all resides in RAM. If swaps occur, it jacks with the cache in sometimes unpredictable ways. As result, it's recommended to turn off swapping completely on the box your node(s) run on, and set Elasticsearch/JVM to have a min and max memory consumption of roughly half the available RAM of the box. That's virtually impossible to achieve if you have other things running on it like a web server or database. Databases in particular aggressively consume RAM in order to increase throughput, which is why those should likewise reside on their own servers.