I use Redis to cache database inserts. For this I created a list CACHE into which I push serialized JSON lists. In pseudocode:
let entries = [{a}, {b}, {c}, ...];
redis.rpush("CACHE", JSON.stringify(entries));
The idea is to run this code for an hour, then later do an
let all = redis.lrange("CACHE", 0, LIMIT);
processAndInsert(all);
redis.ltrim("CACHE", 0, all.length);
Now the thing is that each entries can be relatively large (but far below 512MB / whatever Redis limit I read about). Each of the a, b, c is an object of probably 20 bytes, and entries itself can easily have 100k+ objects / 2MB.
My problem now is that even for very short CACHE lists of only 15 entries a simple lrange can take many minutes(!) even from the redis-cli (my node.js actually dies with an "FATAL ERROR: CALL_AND_RETRY_LAST Allocation failed - process out of memory", but that's a side comment).
The debug output for the list looks like this:
127.0.0.1:6379> debug object "CACHE"
Value at:00007FF202F4E330 refcount:1 encoding:linkedlist serializedlength:18104464 lru:12984004 lru_seconds_idle:1078
What is happening? Why is this so massively slow, and what can I do about it? This does not seem to be a normal slowness, something seems to be fundamentally wrong.
I am using a local Redis 2.8.2101 (x64), ioredis 1.6.1, node.js 0.12 on a relatively hardcore Windows 10 gaming machine (i5, 16GB RAM, 840 EVO SSD, ...) by the way.
Redis is great at doing lots of small operations,
but not so great at doing small numbers of "very big" operations.
I think you should re-evaluate your algorithm, and try to break apart your data in to smaller chunks. Not only you'll save the bandwidth, you'll also will not lock your redis instance long amounts of time.
Redis offers many data structures you should be able to use for more fine grain control over your data.
Well, still, in this case, since you are running the redis locally, and assuming you are not running anything else but this code, I doubt that the bandwidth, nor the redis is the problem. I'm more thinking this line:
JSON.stringify()
is the main culprit why you are seeing the slow execution.
JSON serialization of 20MB of string is not something simple,
The process needs allocate many small strings, and also has to go through all of your array and inspect each item individually. All of this will take a long time for a big object like this one.
Again, if you were breaking apart your data, and doing smaller operations with redis, you'd not need the JSON serializer at all.
Related
I'm wondering the most efficient way to store this data.
I need to track 30-50 million data points per day. It needs to be extremely fast read/write, so I'm using redis.
The data only needs to last for 24 hours, at which point it will EXPIRE.
The data looks like this as a key/value hash
{
"statistics:a5ded391ce974a1b9a86aa5322ea9e90": {
xbi: 1,
bid: 0.24024,
xpl: 25.0,
acc: 40,
pid: 43,
cos: 0.025,
xmp: "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx",
clu: "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
}
}
I've replaced the actual string with a lot of x but that IS the proper length of the string.
So far, according to my calculations.... this will use hundreds of GB of memory. Does that seem correct?
This is mostly ephemeral logging data thats important, but not important enough to try to support writing to disk or failovers. I am comfortable keeping it on 1 machine, if that helps make this easier.
What would be the best way to reduce memory space in this scenario? Is there a better way I can do this? Does redis support 300GB on a single instance?
In redis.conf - set hash-max-ziplist-value to 1 more than the length of the field 'xmp'. Then restart redis, and watch your memory go down significantly.
The default value is 64. Increasing it increases cpu utilization when you modify or add new fields in the hash. But your use case seems to be create-only, and in that case there shouldn't be any drawbacks of increasing the setting.
this will use hundreds of GB of memory. Does that seem correct?
YES
Does redis support 300GB on a single instance?
YES
Is there a better way I can do this?
You can try the following methods:
Avoid Using Hash
Since you always get all fields of the log with HGETALL, there's NO need to save the log as HASH. HASH consumes more memory than STRING.
You can serialize all fields into a string, and save the log as a key-value pair:
SET 'statistics:a5ded391ce974a1b9a86aa5322ea9e90' '{xbi: 1, bid: 0.24024, and other fields}'
#Sripathi Krishnan's answer gives another way to avoid HASH, i.e. config Redis to encode the HASH into ZIPLIST. It's a good idea if you don't share your Redis with other applications. Otherwise, this modification might cause problem to others.
Compress The Data
In order to reduce memory usage, you can try to compress your data. Redis can store binary strings, so you can use gzip, snappy or other compression algorithm to compress the log text into binary string, and save it into Redis.
Normally, you can get better compression when the input is bigger. So you'd better compress the whole log, instead of compress each field one by one.
The side-effect is that the producer and consumer of the log need to cost some CPU to compress and decompress the data. However, normally that's NOT a problem, and also it can reduce some network bandwidth.
Batch Write and Batch Read
As I mentioned above, if you want to get better compression, you should get a bigger input. So if you can write multiple logs in a batch, you can compress the batch of logs to get better compression.
Compress multiple logs into a batch: compress(log1, log2, log3) -> batch1: batch-result
Put the batch result into Redis as a key-value pair: SET batch1 batch-result
Build an index for the batch: MSET log1 batch1 log2 batch1 log3 batch1
When you need to get the log:
Search the index to get the batch key: GET log1 -> batch1
Get the batch result: GET batch1 -> batch-result
Decompress the batch result and look up the log from the result
The last method is the most complicated one, and the extra index will cost some extra memory. However, it can largely reduce the size of your data.
Also what these methods can achieve, largely depends on your log. You should do lots of benchmark :)
I'm going to improve OCL kernel performance and want to clarify how memory transactions work and what memory access pattern is really better (and why).
The kernel is fed with vectors of 8 integers which are defined as array: int v[8], that means, before doing any computation entire vector must be loaded into GPRs. So, I believe the bottleneck of this code is initial data load.
First, I consider some theory basics.
Target HW is Radeon RX 480/580, that has 256 bit GDDR5 memory bus, on which burst read/write transaction has 8 words granularity, hence, one memory transaction reads 2048 bits or 256 bytes. That, I believe, what CL_DEVICE_MEM_BASE_ADDR_ALIGN refers to:
Alignment (bits) of base address: 2048.
Thus, my first question: what is the physical sense of 128-byte cacheline? Does it keep the portion of data fetched by single burst read but not really requested? What happens with the rest if we requested, say, 32 or 64 bytes - thus, the leftover exceeds the cache line size? (I suppose, it will be just discarded - then, which part: head, tail...?)
Now back to my kernel, I think that cache does not play a significant role in my case because one burst reads 64 integers -> one memory transaction can theoretically feed 8 work items at once, there is no extra data to read, and memory is always coalesced.
But still, I can place my data with two different access patterns:
1) contiguous
a[i] = v[get_global_id(0) * get_global_size(0) + i];
(wich actually perfomed as)
*(int8*)a = *(int8*)v;
2) interleaved
a[i] = v[get_global_id(0) + i * get_global_size(0)];
I expect in my case contiguous would be faster because as said above one memory transaction can completely stuff 8 work items with data. However, I do not know, how the scheduler in compute unit physically works: does it need all data to be ready for all SIMD lanes or just first portion for 4 parallel SIMD elements would be enough? Nevertheless, I suppose it is smart enough to fully provide with data at least one CU first, as soon as CU's may execute command flows independently.
While in second case we need to perform 8 * global_size / 64 transactions to get a complete vector.
So, my second question: is my assumption right?
Now, the practice.
Actually, I split entire task in two kernels because one part has less register pressure than another and therefore can employ more work items. So first I played with pattern how the data stored in transition between kernels (using vload8/vstore8 or casting to int8 give the same result) and the result was somewhat strange: kernel that reads data in contiguous way works about 10% faster (both in CodeXL and by OS time measuring), but the kernel that stores data contiguously performs surprisingly slower. The overall time for two kernels then is roughly the same. In my thoughts both must behave at least the same way - either be slower or faster, but these inverse results seemed unexplainable.
And my third question is: can anyone explain such a result? Or may be I am doing something wrong? (Or completely wrong?)
Well, not really answered all my question but some information found in vastness of internet put things together more clear way, at least for me (unlike abovementioned AMD Optimization Guide, which seems unclear and sometimes confusing):
«the hardware performs some coalescing, but it's complicated...
memory accesses in a warp do not necessarily have to be contiguous, but it does matter how many 32 byte global memory segments (and 128 byte l1 cache segments) they fall into. the memory controller can load 1, 2 or 4 of those 32 byte segments in a single transaction, but that's read through the cache in 128 byte cache lines.
thus, if every lane in a warp loads a random word in a 128 byte range, then there is no penalty; it's 1 transaction and the reading is at full efficiency. but, if every lane in a warp loads 4 bytes with a stride of 128 bytes, then this is very bad: 4096 bytes are loaded but only 128 are used, resulting in ~3% efficiency.»
So, for my case it does not realy matter how the data is read/stored while it is always contiguous, but the order the parts of vectors are loaded may affect the consequent command flow (re)scheduling by compiler.
I also can imagine that newer GCN architecture can do cached/coalesced writes, that is why my results are different from those prompted by that Optimization Guide.
Have a look at chapter 2.1 in the AMD OpenCL Optimization Guide. It focuses mostly on older generation cards but the GCN architecture did not completely change, therefore should still apply to your device (polaris).
In general AMD cards have multiple memory controllers to which in every clock cycle memory requests are distributed. If you for example access your values in column-major instead of row-major logic your performance will be worse because the requests are sent to the same memory controller. (by column major I mean a column of your matrix is accessed together by all the work-items executed in the current clock cycle, this is what you refer to as coalesced vs interleaved). If you access one row of elements (meaning coalesced) in a single clock cycle (meaning all work-items access values within the same row), those requests should be distributed to different memory controllers rather than the same.
Regarding alignment and cache line sizes, I'm wondering if this really helps improving the performance. If I were in your situation I would try to have a look whether I can optimize the algorithm itself or if I access the values often and it would make sense to copy them to the local memory. But than again it is hard to tell without any knowledge about what your kernels execute.
Best Regards,
Michael
I have following scenario:
Fetch array of numbers (from REDIS) conditionally
For each number do some async stuff (fetch something from DB based on number)
For each thing in result set from DB do another async stuff
Periodically repeat 1. 2. 3. because new numbers will be constantly added to REDIS structure.Those numbers represent unix timestamp in milliseconds so out of the box those numbers will always be sorted in time of addition
Conditionally means fetch those unix timestamp from REDIS that are less or equal to current unix timestamp in milliseconds(Date.now())
Question is what REDIS data type fit the most for this use case having in mind that this code will be scaled up to N instances, so N instances will share access to single REDIS instance. To equally share the load each instance will read for example first(oldest) 5 numbers from REDIS. Numbers are unique (adding same number should fail silently) so REDIS SET seems like a good choice but reading M first elements from REDIS set seems impossible.
To prevent two different instance of the code to read same numbers REDIS read operation should be atomic, it should read the numbers and delete them. If any async operation fail on specific number (steps 2. and 3.), numbers should be added again to REDIS to be handled again. They should be re-added back to the head not to the end to be handled again as soon as possible. As far as i know SADD would push it to the tail.
SMEMBERS key would read everything, it looks like a hammer to me. I would need to include some application logic to get first five than to check what is less or equal to Date.now() and then to delete those and to wrap somehow everything in single transaction. Besides that set cardinality can be huge.
SSCAN sounds interesting but i don't have any clue how it works in "scaled" environment like described above. Besides that, per REDIS docs: The SCAN family of commands only offer limited guarantees about the returned elements since the collection that we incrementally iterate can change during the iteration process. Like described above collection will be changed frequently
A more appropriate data structure would be the Sorted Set - members have a float score that is very suitable for storing a timestamp and you can perform range searches (i.e. anything less or equal a given value).
The relevant starting points are the ZADD, ZRANGEBYSCORE and ZREMRANGEBYSCORE commands.
To ensure the atomicity when reading and removing members, you can choose between the the following options: Redis transactions, Redis Lua script and in the next version (v4) a Redis module.
Transactions
Using transactions simply means doing the following code running on your instances:
MULTI
ZRANGEBYSCORE <keyname> -inf <now-timestamp>
ZREMRANGEBYSCORE <keyname> -inf <now-timestamp>
EXEC
Where <keyname> is your key's name and <now-timestamp> is the current time.
Lua script
A Lua script can be cached and runs embedded in the server, so in some cases it is a preferable approach. It is definitely the best approach for short snippets of atomic logic if you need flow control (remember that a MULTI transaction returns the values only after execution). Such a script would look as follows:
local r = redis.call('ZRANGEBYSCORE', KEYS[1], '-inf', ARGV[1])
redis.call('ZREMRANGEBYSCORE', KEYS[1], '-inf', ARGV[1])
return r
To run this, first cache it using SCRIPT LOAD and then call it with EVALSHA like so:
EVALSHA <script-sha> 1 <key-name> <now-timestamp>
Where <script-sha> is the sha1 of the script returned by SCRIPT LOAD.
Redis modules
In the near future, once v4 is GA you'll be able to write and use modules. Once this becomes a reality, you'll be able to use this module we've made that provides the ZPOP command and could be extended to cover this use case as well.
Im using datastax-4.6. I have created a cassandra table and stored 2crore records. Im trying to read the data using scala. The code works fine for few records but when i try to retrieve all 2crore records it displays me follwing error.
**WARN BlockManagerMasterActor: Removing BlockManager BlockManagerId(1, 172.20.98.17, 34224, 0) with no recent heart beats: 140948ms exceeds 45000ms
15/05/15 19:34:06 ERROR ConnectionManager: Corresponding SendingConnection to ConnectionManagerId(C15759,34224) not found**
Any help?
This problem is often tied to GC pressure
Tuning your Timeouts
Increase the spark.storage.blockManagerHeartBeatMs so that Spark waits for the GC pause to end.
SPARK-734 recommends setting -Dspark.worker.timeout=30000 -Dspark.akka.timeout=30000 -Dspark.storage.blockManagerHeartBeatMs=30000 -Dspark.akka.retry.wait=30000 -Dspark.akka.frameSize=10000
Tuning your jobs for your JVM
spark.cassandra.input.split.size - will allow you to change the level of parallelization of your cassandra reads. Bigger split sizes mean that more data will have to reside in memory at the same time.
spark.storage.memoryFraction and spark.shuffle.memoryFraction - amount of the heap that will be occupied by RDDs (as opposed to shuffle memory and spark overhead). If you aren't doing any shuffles, you could increase this value. The databricks guys say to make this similar in size to the size of your oldgen.
spark.executor.memory - Obviously this depends on your hardware. Per DataBricks you can do up to 55gb. Make sure to leave enough RAM for C* and for your OS and OS page cache. Remember that long GC pauses happen on larger heaps.
Out of curiosity, are you frequently going to be extracting your entire C* table with Spark? What's the use case?
I used to work with SQL like MySQL, Postgres or MSSQL.
Now I want to play with Redis. I'm working on a little home project, that I think is the best choice for starting using Redis.
I have a machine that reads temperature (indoor and outdoor) and humidity. I need to store the readings into Redis. Can you help me to understand the best data structure to do so?
Other than this data I need to store the time (ex. unix timestamp) of the temperature reading for use plotting a graphic.
I installed Redis read the documentation, so I understand the commands and data types.
Since this is your first Redis project and it's a home project, I'd be careful about being to careful. Here's a couple ways to consider designing it (NOTE: I only dug deep into REDIS this past weekend so hopefully others will weigh in).
IDEA 1:
Four ordered sets
KEY for sets are "indoor_temps", "outdoor_temps", "indoor_humidity", "outdoor_humidity"
VALUES are the temperatures / humidities
SCORE is the date stored as EPOCH
IDEA 2:
Four types of keys (best shown by example)
datetime_key = /year:2014/month:07/day:12/hour:07/minute:32/second:54
type_keys = [indoor_temps, outdoor_temps, indoor_humidity, outdoor_humidity]
keys are of form type + "/" + datetime_key
values are the temp and humidity itself
You probably want to implement some initial design and then work with the data immediately - graph it, do stats, etc. Whatever you plan to do with it. That will expose flaws and if they are major, flush the database and try again. These designs should really only take ~1 hour to implement since the only thing you're really changing is a few Redis commands and some string manipulation to convert the data to keys.
I like Tony's suggestions, but I'll also throw out another possibility.
4 lists
keys are "indoor_temps", "outdoor_temps", "indoor_humidity", "outdoor_humidity"
values are of the form < timestamp >_< reading > ie.( "1403197981_27.2" )
Push items onto the front of the list using LPUSH. Get a set of readings using LRANGE. The list will always be ordered by the time of the reading. Obviously split the value on "_" to get your time and reading...
In all honesty, this will give the same properties as Tony's first example, with slightly worse lookup performance, but better memory usage. I'm guessing for this project you'll be neither memory, nor CPU constrained, so the choice is probably not an issue. That said, if you expect to be saving 100's of thousands or more readings, I would suggest the list unless you want to consume a large portion of your system's memory.
Also, it's a good idea to call EXPIRE on your entries with some reasonable TTL that encompasses the length of time you want to save the readings for. If your plan is to have them live in perpetuity then you may want to look at backing them up to a disk DB over time, and just use Redis as a quick lookup cache for recent readings.
Thank to all answer, I choose this strucure:
4 lists: tempIN, tempOut, humidIN and humidOUT
values are: [value]:[timestamp]. For example: "25.4:1403615247"
As suggested from wallacer i want to backup old entries out from Redis.
For main frontend i need only last two days of sample.
For example i can create Redis RDB file snapshot and "trim" the live lists. This solution is not convenient in the event that, in the future you want to recover old values.
Do you have any tips on what kind of procedure to adopt to store the data? Maybe use of SQLIte DB?