You are browsing a read-only backup copy of Wikitech. The primary site can be found at wikitech.wikimedia.org

Analytics/Data access: Difference between revisions

From Wikitech-static
Jump to navigation Jump to search
imported>Nuria
imported>Neil P. Quinn-WMF
(Update Analytics clients section)
Line 47: Line 47:
!Access to stat1006
!Access to stat1006
!Access to stat1004
!Access to stat1004
!Access to stat1005
!Access to notebook100[34]
!Access to notebook100[34]
|-
|-
Line 52: Line 53:
|
|
|X
|X
|
|
|
|X
|X
Line 58: Line 60:
|
|
|X
|X
|
|
|
|
|
Line 64: Line 67:
|X
|X
|X
|X
|
|
|
|X
|X
Line 70: Line 74:
|X
|X
|
|
|X
|X
|X
|X
|X
Line 76: Line 81:
|X
|X
|
|
|X
|X
|X
|X
|X
Line 81: Line 87:
|<code>analytics-wmde</code>
|<code>analytics-wmde</code>
|X
|X
|
|
|
|
|
Line 145: Line 152:
|X
|X
|
|
|}
|}<br />


==Infrastructure==
==Infrastructure==


=== Analytics clients ===
=== Analytics clients ===
Once you have access to the production cluster, there are several servers which you can use to access the various private data sources and do general statistical computation. There are two types: the stat servers, designed for command-line use, and the [[SWAP]] servers, designed for [https://jupyter.org Jupyter] notebook use. Together, these are called the [[phab:source/operations-puppet/browse/production/modules/profile/manifests/analytics/cluster/client.pp|analytics clients]], since they act as clients accessing data from various other databases.  
{{warning|1= The Analytics clients store data using redundant [[w:en:RAID|RAID]] configurations, but are not otherwise backed up. Your home directory on HDFS (<code>/user/your-username</code>) is a safer place for important data.}}Once you have access to the production cluster, there are several servers which you can use to access the various private data sources and do general statistical computation. There are two types: the stat servers, designed for command-line use, and the [[SWAP]] servers, designed for [https://jupyter.org Jupyter] notebook use. Together, these are called the [[phab:source/operations-puppet/browse/production/modules/profile/manifests/analytics/cluster/client.pp|analytics clients]], since they act as clients accessing data from various other databases.  
{| class="wikitable"
{| class="wikitable"
!Name
!Name
Line 167: Line 174:
|stat1004.eqiad.wmnet
|stat1004.eqiad.wmnet
|Hadoop
|Hadoop
|-
|[[stat1005]]
|stat1005.eqiad.wmnet
|Hadoop, GPU
|-
|-
|notebook1003
|notebook1003
|notebook1003.eqiad.wmnet
|notebook1003.eqiad.wmnet
|Hadoop, MariaDB
|Hadoop, MariaDB, SWAP
|-
|-
|notebook1004
|notebook1004
|notebook1004.eqiad.wmnet
|notebook1004.eqiad.wmnet
|Hadoop, MariaDB
|Hadoop, MariaDB, SWAP
|}You may need to access the internet from the stats machines (for example, to download a Python script using <code>pip</code>). By default, this will fail because the machines are tightly firewalled. You'll have to use the [[Http proxy|HTTP proxy]].
|}You may need to access the internet from the analytics clients (for example, to download a Python script using <code>pip</code>). By default, this will fail because the machines are tightly firewalled. You'll have to use the [[Http proxy|HTTP proxy]].
 
{{warning|1= The home directories of the above hosts are not backed up, but they run on software RAID configurations spanning multiple disks. Please keep this in mind and backup important data if needed. Your home directory on HDFS (/user/your-username) is a safer place for important data that is too big to backup on a regular host/laptop.}}
 
===MariaDB replicas===
The SRE team (Data Persistence) maintains several dedicated MariaDB replicas. These contain copies of the production [[Mw:Manual:Database layout|MediaWiki databases]] (both actively-used mainstream projects and small internal-facing wikis, like various projects' Arbitration Committees) as well as the [[Analytics/EventLogging|EventLogging]] databases.
 
<br />
 
{{warning|1= analytics-store.eqiad.wmnet (dbstore1002) has been deprecated in [[phab:T210478]]}}
 
<br />
 
{{warning|1= The sX-analytics-slave.eqiad.wmnet CNAMEs have been deprecated in favor of using only analytics-store.eqiad.wmnet.}}
 
Up to T210478, all the wiki databases were replicated in one single MySQL instance/host, named <code>analytics-store.eqiad.wmnet</code>, known for short as <code>analytics-store</code> (this hostname was actually an alias for <code>dbstore1002.eqiad.wmnet</code>). Several CNAMEs like <code>x1-analytics-slave.eqiad.wmnet</code> were maintained to ease the access to dbstore1002. This environment has been deprecated in favor of a multi instance/host solution, namely each section (sX or x1) is now running in a separate MySQL database instance. In order to ease the access, the following CNAMEs have been created:
 
* s1-analytics-replica.eqiad.wmnet
* s2-analytics-replica.eqiad.wmnet
* ...
* s8-analytics-replica.eqiad.wmnet
* x1-analytics-replica.eqiad.wmnet
*staging-db-analytics.eqiad.wmnet
 
The main gotcha now is to figure out what port to use to connect to each section, since as we said above there are multiple MySQL instances running. This is the scheme to use:
 
* 331 + the digit of the section in case of sX. Example: s5 will be accessible at s5-analytics-replica.eqiad.wmnet:3315
* 3320 for x1. Example: x1-analytics-replica.eqiad.wmnet:3320
* 3350 for staging
 
We have created DNS [[:en:SRV_record|SRV]] records to ease the use of the dbstore shard into scripts (see Python functions below for an example):
 
* _s1-analytics._tcp.eqiad.wmnet
* ...
* _s8-analytics._tcp.eqiad.wmnet
* _x1-analytics._tcp.eqiad.wmnet
* _staging-analytics._tcp.eqiad.wment
 
These records look "weird" if you are not familiar with them, but they allow us to get a hostname/port combination from one DNS name! <u>Please note: the mysql client is sadly not capable of reading those, so you cannot use them directly with it.</u> 
 
You can access these analytics replicas from either stat1007 or stat1006. To use the <code>mysql</code> client interactively, type: <code>mysql -h sX-analytics-replica.eqiad.wmnet -P 331X -A</code>. The <code>-A</code> disables tab autocompletion, which can be bothersome when pasting multi-line queries into the prompt. You'll then be dropped into the MySQL command line.<ref>The stat machines automatically authenticate to the MariaDB replicas using passwords stored in preloaded config (<code>.cnf</code>) files. On stat1006, the file is <code>/etc/mysql/conf.d/research-client.cnf</code>, while on stat1007 it's <code>/etc/mysql/conf.d/analytics-research-client.cnf</code>. These files are automatically referenced from <code>/etc/mysql/my.cnf</code>, which the command line <code>mysql</code> program reads by default. Other programs may not do this automatically, and require an explicit pointer to the underlying <code>.cnf</code> file.</ref>
 
If you'd rather generate a TSV file<ref>The <code>mysql</code> utility doesn't have the ability to generate files in other formats like CSV.</ref> and then retrieve it later, you can also do so from the command line. You can type type:
 
<code>mysql -h s1-analytics-replica.eqiad.wmnet -P 3321 <nowiki>{{database name}}</nowiki> -e "<nowiki>{{your query;}}</nowiki>" > <nowiki>{{filename}}</nowiki>.tsv</code>
 
It'll go off to generate the file on its own.<ref>The file extension you choose doesn't actually affect the command, but since <code>-e</code> generates a TSV file, you should use the corresponding file extension so other programs load the resulting file correctly.</ref>
 
As well as connecting directly, it's also possible to connect automatically from your programming language of choice, be it R or Python. For Python, we have the [http://mysql-python.sourceforge.net/MySQLdb.html MySQLdb] module installed on stat1006 and stat1007. For R, we have [http://cran.r-project.org/web/packages/RMySQL/RMySQL.pdf RMySQL].
 
The following Python snippets are an example about how it is possible to find the correct Mysql hostname+port combination given a wiki db name:<syntaxhighlight lang="python3">
# The second function needs dnspython to work
import dns.resolver
 
def get_mediawiki_section_dbname_mapping(mw_config_path, use_x1):
    db_mapping = {}
    if use_x1:
        dblist_section_paths = [mw_config_path.rstrip('/') + '/dblists/all.dblist']
    else:
        dblist_section_paths = glob.glob(mw_config_path.rstrip('/') + '/dblists/s[0-9]*.dblist')
    for dblist_section_path in dblist_section_paths:
        with open(dblist_section_path, 'r') as f:
            for db in f.readlines():
                db_mapping[db.strip()] = dblist_section_path.strip().rstrip('.dblist').split('/')[-1]
 
    return db_mapping
 
 
def get_dbstore_host_port(db_mapping, use_x1, dbname):
    if dbname == 'staging':
        shard = 'staging'
    elif use_x1:
        shard = 'x1'
    else:
        try:
            shard = db_mapping[dbname]
        except KeyError:
            raise RuntimeError("The database {} is not listed among the dblist files of the supported sections."
                              .format(dbname))
    answers = dns.resolver.query('_' + shard + '-analytics._tcp.eqiad.wmnet', 'SRV')
    host, port = str(answers[0].target), answers[0].port
    return (host,port)
</syntaxhighlight>Example of a Python script to retrieve data for a specific wiki from a Notebook (credits to Neil for the work!):<syntaxhighlight lang="python">
import dns.resolver
import mysql.connector as mysql
 
 
query = "select count(*) from recentchanges"
ans = dns.resolver.query('_s7-analytics._tcp.eqiad.wmnet', 'SRV')[0]
conn = mysql.connect(
    host=str(ans.target),
    port=ans.port,
    database="hewiki",
    option_files='/etc/mysql/conf.d/research-client.cnf',
    charset='binary',
    autocommit=False
)
cursor = conn.cursor()
cursor.execute(query)
cursor.fetchall()
</syntaxhighlight>
This is of course a quick solution, it can be done in any number of ways. If you have comments/suggestions/improvements please feel free to improve :)
 
The first function, get_mediawiki_section_dbname_mapping, needs the path of the [[gerrit:#/admin/projects/operations/mediawiki-config|mediawiki-config repository]] local checkout. '''If you run your script on a stat100* or notebook100* host, then you can find the repository checked out under /srv/mediawiki-config'''. This is one source of the mapping between mediawiki db sections (s1..s8, x1) to wiki names. For example, itwiki is contained into two files:
 
* dblists/s2.dblist
* dblists/all.dblist
 
The above files means two things:
 
1) itwiki's database will be available on s2-analytics-replica.eqiad.wmnet
 
2) itiwiki's database will be available on x1-analytics-replica.eqiad.wmnet (if we need the tables related to the extensions, since x1 contains all the wikis).
 
This is a pre-requisite to use the get_dbstore_host_port, that starts from the mapping described above and uses DNS SRV records as explained above.
 
====MySQL wrapper====
On all the Analytics nodes (stat100[4,6,7] and notebook100[3,4]) there is a tool called <code>analytics-mysql</code> that should help using the new databases:<syntaxhighlight lang="bash">
elukey@stat1006:~$ analytics-mysql itwiki --print-target
dbstore1004.eqiad.wmnet:3312
 
elukey@stat1006:~$ analytics-mysql itwiki --print-target --use-x1
dbstore1005.eqiad.wmnet:3320
 
elukey@stat1006:~$ analytics-mysql itwiki
Reading table information for completion of table and column names
You can turn off this feature to get a quicker startup with -A
 
Welcome to the MariaDB monitor.  Commands end with ; or \g.
Your MariaDB connection id is 465860
Server version: 10.1.37-MariaDB MariaDB Server
 
Copyright (c) 2000, 2017, Oracle, MariaDB Corporation Ab and others.


Type 'help;' or '\h' for help. Type '\c' to clear the current input statement.
==== Ad-hoc datasets ====
If you wish to '''publish''' a dataset or report from one of the analytics clients, you can place it in the <code>/srv/published-datasets</code>.  You can find more information on Analytics/Ad hoc datasets.


mysql:research@dbstore1004.eqiad.wmnet. [itwiki]> Bye
==== GPU usage ====
On stat1005 we have deployed an AMD GPU for [[phab:T148843|T148843]]. The long term plan is to make it available for all the users logging in, but for the moment its access is restricted to the POSIX group <code>gpu-testers</code> to better test it (and avoid usage contention etc..). Please reach out to the Analytics team if you wish to get added to the group to test the GPU for your use case.


elukey@stat1006:~$ analytics-mysql itwiki -e 'show tables limit 3'
===MariaDB===
ERROR 1064 (42000) at line 1: You have an error in your SQL syntax; check the manual that corresponds to your MariaDB server version for the right syntax to use near 'limit 3' at line 1
The [[Analytics/Systems/MariaDB|Analytics MariaDB cluster]] contains copies of the production [[Mw:Manual:Database layout|MediaWiki databases]] (both actively-used mainstream projects and small internal-facing wikis, like various projects' Arbitration Committees) as well as the [[Analytics/EventLogging|EventLogging]] databases.
elukey@stat1006:~$ analytics-mysql itwiki -e 'show tables'
+--------------------------+
| Tables_in_itwiki        |
+--------------------------+
[..]
</syntaxhighlight>


===Hadoop===
=== Hadoop ===
Finally, we have Hadoop - our storage system for large amounts of data. The easiest way to query the Hadoop data is through [[Analytics/Cluster/Hive|Hive]], which can be accessed from [[stat1005|stat1007]] and stat1004 -  simply type <code>beeline</code> in the terminal, switch to the <code>wmf</code> database, and input your query.
Finally, we have Hadoop - our storage system for large amounts of data. The easiest way to query the Hadoop data is through [[Analytics/Cluster/Hive|Hive]], which can be accessed from [[stat1005|stat1007]] and stat1004 -  simply type <code>beeline</code> in the terminal, switch to the <code>wmf</code> database, and input your query.



Revision as of 12:16, 25 September 2019

In addition to a variety of publicly-available data sources, Wikimedia has a parallel set of private data sources. The main reason is to allows a carefully vetted set of users to perform research and analysis on confidential user data (such as the IP addresses of readers and editors) which is stored according to our privacy policy and data retention guidelines. This private infrastructure also provides duplicate copies of publicly-available data for ease of use.

Access

This private data lives in same server cluster that runs Wikimedia's production websites. This means you will need production shell access to get it (see also these notes on configuring SSH specifically for the purpose of working with the stats servers).

However, since this access gets you closer to both those production websites and this confidential data, it is not freely given out. First, you have to demonstrate a need for these resources. Second, you need to have a non-disclosure agreement with the Wikimedia Foundation. If you're a Foundation employee, this was included as part of your employment agreement. If you're a researcher, it's possible to be sponsored through a formal collaboration with the Wikimedia Foundation's Research team.

User responsibilities

If you get this access, you must remember that this access is extremely sensitive. You have a duty to protect the privacy of our users. As Uncle Ben says, "with great power comes great responsibility." Always follow the rules outlined in the Acknowledgement of Server Access Responsibilities, which you have signed if you have access to this data.

In addition, keep in mind the following important principles:

  • Be paranoid about personally identifiable information (PII). Familiarize yourself with the data you are working on, and determine if it contains any PII. It's better to double and triple check than to assume anything, but if you have any doubt ask the Analytics team (via IRC or email or Phabricator). Please see the data retention guidelines.
  • Don't copy sensitive data (for example, data accessible only by the users in the analytics-privatedata-users) from its origin location to elsewhere (in HDFS or on any other host/support) unless strictly necessary. And most importantly, do it only if you know what you are doing. If you are in doubt, please reach out to the Analytics team first.
  • Restrict access. If you do need to copy sensitive data somewhere, please make sure that you are the only one able to access the data. For example, if you copy Webrequest data from its location on HDFS to your /user/$your-username directory, make sure that the permissions are set to avoid everybody with access to HDFS to read the data. This is essential to avoid accidental leaks of PII/sensitive data or retention over our guidelines (https://meta.wikimedia.org/wiki/Data_retention_guidelines).
  • Clean up copies of data. Please make sure that any data that you copied is deleted as soon as your work has been done.

If you ever have any questions or doubts, err on the side of caution and contact the Analytics team. We are very friendly and happy to help!

Access Groups

To get access, you submit a request on phabricator and tag SRE-Access-Requests. You will need to specify what access group you need.

'analytics-*' groups have access to the Analytics Cluster (which mostly means Hadoop). 'statistics-*' groups get access to stat* servers for local (non distributed) compute resources. These groups overlap in what servers they grant ssh access to, but further posix permissions restrict access to things like MySQL, Hadoop, and files.

Here's a summary of groups you might need (as of 2016-10-18):

researchers
Access to stat1006 and the credentials for the MariaDB slaves in/etc/mysql/conf.d/research-client.cnf.
statistics-users
Access to stat1006 for number crunching and accessing non private log files hosted there.
statistics-privatedata-users
Access to stat100[56], public data like sampled Webrequest logs (stored under /a/log/webrequest/archive) and for the MariaDB slaves in /etc/mysql/conf.d/statistics-private-client.cnf
analytics-wmde
WMDE specific group (mostly used for crons). Access to stat1007 and to MariaDB slaves in /etc/mysql/conf.d/research-wmde-client.cnf
analytics-users
Access to stat1004 to connect to the Analytics/Cluster (Hadoop/Hive) (NO HADOOP PRIVATE DATA).
analytics-privatedata-users
Access to stat1007 and stat1004 to connect to the Analytics/Cluster (Hadoop/Hive) and to query private data hosted there, including webrequest logs. Access to MariaDB slaves in /etc/mysql/conf.d/analytics-research-client.cnf
If you want Hadoop access you probably want this.

The list of users currently in each group is available in this configuration file.[1]

Host access granted

Access Groups Access to stat1007 Access to stat1006 Access to stat1004 Access to stat1005 Access to notebook100[34]
researchers X X
statistics-users X
statistics-privatedata-users X X X
analytics-users X X X X
analytics-privatedata-users X X X X
analytics-wmde X

Data access granted

Access Groups Hadoop access

(No private data)

Hadoop access

(Private data)

research-client.cnf statistics-private-client.cnf research-wmde-client.cnf analytics-research-client.cnf
researchers X
statistics-users
statistics-privatedata-users X
analytics-users X
analytics-privatedata-users X X X
analytics-wmde X


Infrastructure

Analytics clients

Once you have access to the production cluster, there are several servers which you can use to access the various private data sources and do general statistical computation. There are two types: the stat servers, designed for command-line use, and the SWAP servers, designed for Jupyter notebook use. Together, these are called the analytics clients, since they act as clients accessing data from various other databases.

Name Hostname Access available
stat1007 stat1007.eqiad.wmnet Hadoop, MariaDB
stat1006 stat1006.eqiad.wmnet MariaDB
stat1004 stat1004.eqiad.wmnet Hadoop
stat1005 stat1005.eqiad.wmnet Hadoop, GPU
notebook1003 notebook1003.eqiad.wmnet Hadoop, MariaDB, SWAP
notebook1004 notebook1004.eqiad.wmnet Hadoop, MariaDB, SWAP

You may need to access the internet from the analytics clients (for example, to download a Python script using pip). By default, this will fail because the machines are tightly firewalled. You'll have to use the HTTP proxy.

Ad-hoc datasets

If you wish to publish a dataset or report from one of the analytics clients, you can place it in the /srv/published-datasets. You can find more information on Analytics/Ad hoc datasets.

GPU usage

On stat1005 we have deployed an AMD GPU for T148843. The long term plan is to make it available for all the users logging in, but for the moment its access is restricted to the POSIX group gpu-testers to better test it (and avoid usage contention etc..). Please reach out to the Analytics team if you wish to get added to the group to test the GPU for your use case.

MariaDB

The Analytics MariaDB cluster contains copies of the production MediaWiki databases (both actively-used mainstream projects and small internal-facing wikis, like various projects' Arbitration Committees) as well as the EventLogging databases.

Hadoop

Finally, we have Hadoop - our storage system for large amounts of data. The easiest way to query the Hadoop data is through Hive, which can be accessed from stat1007 and stat1004 - simply type beeline in the terminal, switch to the wmf database, and input your query.

At the moment there are no recommended Hive access packages for R or Python. In the meantime, the best way to get data out of the system is to treat it as you would the Analytics slaves; through the terminal, type:

beeline -f my_query.hql > file_name.tsv

For information about writing HQL to query this data, see the Hive language manual.

Data sources

Data sets and data streams can be found in Category:Data_stream

Data Dashboards. Superset and Turnilo (previously called Pivot)

Superset: http://superset.wikimedia.org Pivot: http://pivot.wikimedia.org

You need a wikitech login that is in the "wmf" or "nda" LDAP groups. If you don't have it, please create a task like https://phabricator.wikimedia.org/T160662

Before requesting access, please make sure you:

Depending on the above, you can request to be added to the wmf group or the nda group. Please indicate the motivation on the task about why you need access and ping the analytics team if you don't hear any feedback soon from the Opsen on duty.

MediaWiki application data

You can do a lot of work with the data stored by MediaWiki in the normal course of running itself. This includes data about:

  • Users' edit counts (consult the user table)
  • Edits to a particular page (consult the revision table, joined with the page table if necessary)
  • Account creations (consult the logging table)

Databases

You can access this data using the replica MariaDB databases. These are accessible from the stat100* machines, as detailed below.

For an overview of how the data is laid out in those databases, consult the database layout manual.

There are a few things that aren't available from the databases replicas. The main example of this is the actual content of pages and revisions. Instead, you can access them through the API or in the XML dumps, which are both described below.

API

A subset of this application data, which doesn't present privacy concerns, is also publicly accessible through the API (except for private wikis, which you shouldn't really need to perform research on anyway!). A good way to understand it, and to test queries, is Special:ApiSandbox, which provides a way of easily constructing API calls and testing them. The output includes "Request URL" - a direct URL for making that query in the future, that should work on any and all Wikimedia production wikis.

If you're interested in common API tasks, and don't feel like reinventing the wheel, there are a number of Python-based API wrappers and MediaWiki utilities. Our very own Aaron Halfaker maintains MediaWiki Utilities, which includes a module dedicated to API interactions. There's no equivalent for R yet.

Database dumps

Every month, XML snapshots of the databases are generated. Since they're generated monthly, they're always slightly outdated, but make up for it by being incredibly cohesive (and incredibly large). They contain both the text of each revision of each page, and snapshots of the database tables. As such, they're a really good way of getting large amounts of diffs or information on revisions without running into the query limits on the API.

Aaron's MediaWiki-utilities package contains a set of functions for handling and parsing through the XML dumps, which should drastically simplify dealing with them. They're also stored internally, as well as through dumps.wikimedia.org, and can be found in /mnt/data/xmldatadumps/public on stat1007.

EventLogging data

One analytics-specific source of data is EventLogging. This allows us to track things we're interested in as researchers that MediaWiki doesn't normally log. Examples include:

  1. A log of changes to user preferences;
  2. A/B testing data;
  3. Clicktracking data.

These datasets are stored in the log database on analytics-slave.eqiad.wmnet (a CNAME for db1108). The schemas that set out each table, and what they contain, can be found on Meta in the Schema namespace.

Pageviews data

An important piece of community-facing data is information on our pageviews; what articles are being read, and how much? This is currently stored in our Hadoop cluster, which contains aggregated pageview data as well as the mostly-raw database of web requests. See the detailed documentation here.

Turnilo

Analytics/Systems/Turnilo-Pivot#Access

Geolocation data

When you have IP addresses - be they from the RequestLogs, EventLogging or MediaWiki itself - you can do geolocation. This can be a very useful way of understanding user behaviour and evaluating how our ecosystem works. We currently use the MaxMind geolocation services, which are accessible on both stat1006 and stat1007: a full guide to geolocation and some examples of how to do it can be found on the 'geolocation' page.

Notes

  1. Other groups including statistics-admins, analytics-admins, eventlogging-admins, and statistics-web-users are for people doing system maintenance and administration, so you don't need them just to access data.