I am running a dataflow job, which readed from BigQuery and scans around 8 GB of data and result in more than 50,000,000 records. Now at group by step I want to group based on a key and one column need to be concatenated . But After concatenated size of concatenated column becomes more than 100 MB that why I have to do that group by in dataflow job because that group by can not be done in Bigquery level due to row size limit of 100 MB.
Now the dataflow job scales well when reading from BigQuery but stuck at Group by Step , I have 2 version of dataflow code, but both are stucking at group by step. When I checked the stack driver logs, it says, processing stuck at lull for more than 1010 sec time(similar kind of message) and Refusing to split GroupedShuffleReader <dataflow_worker.shuffle.GroupedShuffleReader object at 0x7f618b406358> kind of message
I expect the group by state to be completed within 20 mins but is stuck for more than 1 hours and never gets finished
I figured out the thing myself.
Below are the 2 changes that I did in my pipeline:
1. I added a Combine function just after the Group by Key, see screenshot
since the Group by key when running on multiple worker, does a lot of network traffic exchange, and by default the network we use, does not allow the inter network communication, so I have to create a firewall rule to allow traffic from one worker to another worker i.e. ip range to network traffic.
Related
I am trying to optimize my Apache Beam pipeline on Google Cloud Platform Dataflow.
Background information: I am trying to read streaming data from PubSub Messages, and aggregate them based on 3 time windows: 1 min, 5 min and 60 min. Such aggregations consists of summing, averaging, finding the maximum or minimum, etc. For example, for all data collected from 1200 to 1201, I want to aggregate them and write the output into BigTable's 1-min column family. And for all data collected from 1200 to 1205, I want to similarly aggregate them and write the output into BigTable's 5-min column. Same goes for 60min.
The current approach I took is to have 3 separate dataflow jobs (i.e. 3 separate Beam Pipelines), each one having a different window duration (1min, 5min and 60min). See https://beam.apache.org/releases/javadoc/2.0.0/org/apache/beam/sdk/transforms/windowing/Window.html. And the outputs of all 3 dataflow jobs are written to the same BigTable, but on different column families. Other than that, the function and aggregations of the data are the same for the 3 jobs.
However, this seems to be very computationally inefficient, and cost inefficient, as the 3 jobs are essentially doing the same function, with the only exception being the window time duration and output column family.
Some challenges and limitations we faced was that from the Apache Beam documentation, it seems like we are unable to create multiple windows of different periods in a singular dataflow job. Also, when we write the final data into big table, we would have to define the table, column family, column, and rowkey. And unfortunately, the column family is a fixed property (i.e. it cannot be redefined or changed given the window period).
Hence, I am wondering if there is a way to only use 1 dataflow job (i.e. 1 Apache Beam pipeline) that fulfils the objective of this project? Which is to aggregate data on different window periods, and write them to different column families of the same BigTable.
I was considering using Split stream: first window by 1-min, then split into 3 streams (1 write to bigtable for 1-min interval, another for 5-min aggregation, and another for 60-min aggregation). However, the problem is that we are working with streaming data and not batch data.
Thank you
I am trying to track server uptimes using redis.
So the approach I have chosen is as follows:
server xyz will keep on sending my service ping indicating that it was alive and working in the last 30 seconds.
My service will store a list of all time intervals during which the server was active. This will be done by storing a list of {startTime, endTime} in redis, with key as name of the server (xyz)
Depending on a user query, I will use this list to generate server uptime metrics. Like % downtime in between times (T1, T2)
Example:
assume that the time is T currently.
at T+30, server sends a ping.
xyz:["{start:T end:T+30}"]
at T+60, server sends another ping
xyz:["{start:T end:T+30}", "{start:T+30 end:T+60}"]
and so on for all pings.
This works fine , but an issue is that over a large time period this list will get a lot of elements. To avoid this currently, on a ping, I pop the last element of the list, check if it can be merged with the latest time interval. If it can be merged, I coalesce and push a single time interval into the list. if not then 2 time intervals are pushed.
So with this my list becomes like this after step 2 : xyz:["{start:T end:T+60}"]
Some problems I see with this approach is:
the merging is being done in my service, and not redis.
incase my service is distributed, The list ordering might get corrupted due to multiple readers and writers.
Is there a more efficient/elegant way to handle this , like maybe handling merging of time intervals in redis itself ?
We're benchmarking our app with different scales of an Azure SQL database, and we're having a hard time saturating the db. Among other things, we've executed this query:
SELECT *
FROM sys.dm_os_wait_stats
ORDER BY wait_time_ms DESC
The top row of the result was something like
wait_type waiting_tasks_count wait_time_ms max_wait_time_ms signal_wait_time_ms
VDI_CLIENT_OTHER 19560 409007428 60016 37281
What is this wait time? What exactly have we been waiting for during those 409000 seconds (almost 5 days)? Google doesn't seem to know what VDI_CLIENT_OTHER is.
VDI_CLIENT_OTHER is used in case of new replica seeding or any other user initiated workflow that triggers copies like update service tier and setting up geo relationship link. High wait time It likely just means we did seeding and the task remained running waiting for additional work items which aren’t arriving.
Looking for some advice on how best to architect/design and build our pipeline.
After some initial testing, we're not getting the results that we were expecting. Maybe we're just doing something stupid, or our expectations are too high.
Our data/workflow:
Google DFP writes our adserver logs (CSV compressed) directly to GCS (hourly).
A day's worth of these logs has in the region of 30-70 million records, and about 1.5-2 billion for the month.
Perform transformation on 2 of the fields, and write the row to BigQuery.
The transformation involves performing 3 REGEX operations (due to increase to 50 operations) on 2 of the fields, which produces new fields/columns.
What we've got running so far:
Built a pipeline that reads the files from GCS for a day (31.3m), and uses a ParDo to perform the transformation (we thought we'd start with just a day, but our requirements are to process months & years too).
DoFn input is a String, and its output is a BigQuery TableRow.
The pipeline is executed in the cloud with instance type "n1-standard-1" (1vCPU), as we think 1 vCPU per worker is adequate given that the transformation is not overly complex, nor CPU intensive i.e. just a mapping of Strings to Strings.
We've run the job using a few different worker configurations to see how it performs:
5 workers (5 vCPUs) took ~17 mins
5 workers (10 vCPUs) took ~16 mins (in this run we bumped up the instance to "n1-standard-2" to get double the cores to see if it improved performance)
50 min and 100 max workers with autoscale set to "BASIC" (50-100 vCPUs) took ~13 mins
100 min and 150 max workers with autoscale set to "BASIC" (100-150 vCPUs) took ~14 mins
Would those times be in line with what you would expect for our use case and pipeline?
You can also write the output to files and then load it into BigQuery using command line/console. You'd probably save some dollars of instance's uptime. This is what I've been doing after running into issues with Dataflow/BigQuery interface. Also from my experience there is some overhead bringing instances up and tearing them down (could be 3-5 minutes). Do you include this time in your measurements as well?
BigQuery has a write limit of 100,000 rows per second per table OR 6M/per minute. At 31M rows of input that would take ~ 5 minutes of just flat out writes. When you add back the discrete processing time per element & then the synchronization time (read from GCS->dispatch->...) of the graph this looks about right.
We are working on a table sharding model so you can write across a set of tables and then use table wildcards within BigQuery to aggregate across the tables (common model for typical BigQuery streaming use case). I know the BigQuery folks are also looking at increased table streaming limits, but nothing official to share.
Net-net increasing instances is not going to get you much more throughput right now.
Another approach - in the mean time while we work on improving the BigQuery sync - would be to shard your reads using pattern matching via TextIO and then run X separate pipelines targeting X number of tables. Might be a fun experiment. :-)
Make sense?
I am considering implementing a Lambda Architecture in order to process events transmitted by multiple devices.
In most cases (averages etc.) its seems to fit my requirements. However, I am stuck trying to model a specific use case. In short...
Each device has a device_id. Every device emits 1 event per second. Each event has an event_id ranging from {0-->10}.
An event_id of 0 indicates START & an event_id of 10 indicates END
All the events between START & END should be grouped into one single group (event_group).
This will produce tuples of event_groups i.e. {0,2,2,2,5,10}, (0,4,2,7,...5,10), (0,10)
This (event_group) might be small i.e. 10 minutes or very large say 3hours.
According to Lambda Architecture these events transmitted by every device are my "Master Data Set".
Currently, the events are sent to HDFS & Storm using Kafka (Camus, Kafka Spout).
In the Streaming process I group by device_id, and use Redis to maintain a set of incoming events in memory, based on a key which is generated each time an event_id=0 arrives.
The problem lies in HDFS. Say I save a file with all incoming events every hour. Is there a way to distinguish these (group_events)?
Using Hive I can group tuples in the same manner. However, each file will also contain "broken" event_groups
(0,2,2,3) previous computation (file)
(4,3,) previous computation (file)
(5,6,7,8,10) current computation (file)
so that I need to merge them based on device_id into (0,2,2,3,4,3,5,6,7,8,10) (multiple files)
Is a Lambda Architecture a fit for this scenario? Or should the streaming process be the only source of truth? I.e. write to hbase, hdfs itself won't this affect the overall latency.
As far as I understand your process, I don't see any issue, as the principle of Lambda Architecure is to re-process regularly all your data on a batch mode.
(by the way, not all your data, but a time frame, usually larger than the speed layer window)
If you choose a large enough time window for your batch mode (let's say your aggregation window + 3 hours, in order to include even the longest event groups), your map reduce program will be able to compute all your event groups for the desired aggregation window, whatever file the distincts events are stored (Hadoop shuffle magic !)
The underlying files are not part of the problem, but the time windows used to select data to process are.