Subscribe: Internet Super Hero Comments
http://blog.ulf-wendel.de/?feed=comments-rss2
Added By: Feedage Forager Feedage Grade C rated
Language: English
Tags:
comment mysql  comment  fabric  mysql fabric  mysql  mysqlnd  pecl mysqlnd  pecl  plugin mysql  sql  ulf wendel  ulf  wendel 
Rate this Feed
Rate this feedRate this feedRate this feedRate this feedRate this feed
Rate this feed 1 starRate this feed 2 starRate this feed 3 starRate this feed 4 starRate this feed 5 star

Comments (0)

Feed Details and Statistics Feed Statistics
Preview: Internet Super Hero Comments

Comments for Ulf Wendel





Last Build Date: Tue, 21 Jun 2016 18:56:26 +0000

 



Comment on PHP mysqlnd memory optimizations: from 49MB to 2MB by Radek

Tue, 21 Jun 2016 18:56:26 +0000

Great read. Thanks for clearing a lot of the internals up for me. Btw, I think there is a typo in the "counter example code" You have while ($row = $res->fetch_all(MYSQLI_ASSOC)) { ; } should probably be... while ($row = $res->fetch_assoc()) { ; }



Comment on MySQL 5.7 – HTTP Plugin for MySQL by MySQL 5.7 – HTTP Plugin for MySQL | Brainiac

Wed, 27 Jan 2016 09:32:09 +0000

[…] http://blog.ulf-wendel.de/2014/mysql-5-7-http-plugin-mysql/ […]



Comment on Not only SQL injection: I don’t trust you! by MySQL-Injection mal anders | David Müller: Webarchitektur

Mon, 11 Jan 2016 21:41:07 +0000

[…] Original entdeckt von ulf-wendel.de. […]



Comment on Supercharging PHP MySQL applications using the best API by Cgpme

Mon, 21 Dec 2015 00:24:10 +0000

I put the slides on the site. Sorry for the ovgisreht.For the future: I'd like to get this project done.My usage is that the BDB XA API is a particular C binding for an implementation of the XA specification. So the API was developed by sleepycat, and XA itself is specified by The Open Group. I could be convinced that some other usage is better.



Comment on MySQL 5.7 – HTTP Plugin for MySQL by PHP-CRUD-API now supports PostgreSQL 9 | WebSetNet

Sun, 08 Nov 2015 20:44:35 +0000

[…] in only a few hundred lines of PHP code. It is quite comparable to the REST functionality of the experimental HTTP plugin in MySQL […]



Comment on MySQL 5.7 – HTTP Plugin for MySQL by PHP-CRUD-API now supports PostgreSQL 9 - LeaseWeb labs

Sun, 18 Oct 2015 11:28:06 +0000

[…] in only a few hundred lines of PHP code. It is quite comparable to the REST functionality of the experimental HTTP plugin in MySQL […]



Comment on MySQL 5.7 – HTTP Plugin for MySQL by Sven

Tue, 13 Oct 2015 10:34:14 +0000

Hi, I am testing the MySQL HTTP plugin. It is working fine so far for standrad HTTP connections. But how to add HTTPS. I have a SSL/TLS certificate in one pem file (like the one included in the plugin). openssl verify my_file.pem positivly verifies the file. But still the error.log states: "SSL error: SSL context is not usable without certificate and private key" Any ideas what I am doing wrong?



Comment on Webinar tomorrow: Top 10 MySQL Tips and Mistakes for PHP Developers by shane

Sat, 10 Oct 2015 07:29:59 +0000

Hi, I have been searching for your webinar but most links lead to know where. There is a lack of good php security stuff on the web, or maybe there is too much security stuff and a lot of it not trust worthy. I am really interested to see what you had to say. TIA shane



Comment on MySQL 5.7 Fabric: any good? by Sheety

Thu, 01 Oct 2015 17:48:15 +0000

Thanks Liran :)I believe I did touch the seucbjt of 2nd-gen sharding technologies when I referred to shard-able drivers and “shardware” (which stands for sharding middleware). My point, however, is that regardless the use of these technologies and no matter how transparent these are, at the end of the day there may be limitations (e.g., cross-shard TX or DB-embedded logic) that stem from the fact that sharding is external to the DBMS.As for MySQL becoming the bottleneck, this is certainly the case but ONLY when considering the standard, SMP-based setup. A distributed system design (such as ours) allows you not only to have the resources independently assigned to each storage partition but also to scale out and parallelize the database engine layer.



Comment on Using MySQL Fabric from any programming language by DZ

Sat, 26 Sep 2015 16:32:12 +0000

Hi, I was working on the Nodejs Connector for mysql fabric. Just came up with the basic connector which currently support only Range Based Sharding. https://github.com/daudzaidi/mysqlfabric-nodejs-connectors Any comments/suggestions would be highly appreciated.



Comment on MySQL 5.7 – HTTP Plugin for MySQL by neo

Sun, 13 Sep 2015 23:55:22 +0000

We can resolve this in two ways. 1. Use it form internal web applications where security doesn't count. or 2. Apply row level access in MySQL and allow users to read/write on whatever rows they have created. - I feel the second way is promising, All you need to do is @ the time of Sign Up logic you just would be creating a legit MySQL user, And keep giving GRANT access for all the row he creates rows also for all the foreign key rows.



Comment on MySQL 5.7: SQL functions for JSON by Choosing the right MySQL High Availability Solution – webinar replay | MySQL High Availability

Thu, 18 Jun 2015 19:49:06 +0000

[…] Is there a storage engine to handle JSON documents? Here’s an interesting blog on that topic […]



Comment on PoC: Using a Group Communication System (Isis2) to improve MySQL Replication HA by Ken Birman

Tue, 27 Jan 2015 01:14:40 +0000

interesting post! I'm actually doing a port of Isis2 to C++ and we may also look at dropping it into hardware -- embedding it in the NIC of an RDMA card that can also do reliable multicast at optical line speeds. Drop me a note if interested. The only thing is that this port is a pain and not going particularly quickly. I have to teach in the spring, and my class is huge, and I'm also finishing some other projects. But I expect to get it done by early summer. It would be fun to work with you folks on this replication solution for MySQL. Where are the bottlenecks? What should I optimize the beejezus out of? -- Ken



Comment on PoC: Using a Group Communication System (Isis2) to improve MySQL Replication HA by shengdi

Mon, 26 Jan 2015 18:22:38 +0000

Hi there, I have another question. For you C++ plugin code, Where can I find the include file "isis2.h" Thanks.



Comment on PoC: Using a Group Communication System (Isis2) to improve MySQL Replication HA by shengdi

Mon, 26 Jan 2015 15:08:32 +0000

Hi there I am a newbi to MySQL. May i ask some questions about the version of MySQL? Which version of MySQL is used for your C++ code? Do I need a MySQL group replication version of MySQL? Or any tutorials and useful links will be appreciated. Thanks a lot



Comment on MySQL 5.7: SQL functions for JSON by admin

Thu, 22 Jan 2015 19:59:42 +0000

No (not yet).



Comment on MySQL 5.7: SQL functions for JSON by Hans

Thu, 22 Jan 2015 14:27:20 +0000

Can the JSON documents be indexed to improve query speed?



Comment on MySQL 5.7 – HTTP Plugin for MySQL by gman

Tue, 02 Dec 2014 18:03:05 +0000

Ulf, Thanks for starting this plug in. This is potentially very useful and long overdue imho. A few comments: 1. the /sql end points must be optional. 2. the /crud/db/table endpoint must return collections of rows. 3. the /crud/db/table endpoint must also support basic where clauses, like parse.com. You posted the source. What's the best way to submit pull requets? Thanks.



Comment on PoC: Using a Group Communication System (Isis2) to improve MySQL Replication HA by PECL/mysqlnd_ms needs updates for MySQL Group Replication | Ulf WendelUlf Wendel

Thu, 30 Oct 2014 00:42:03 +0000

[…] could make PECL/mysqlnd_ms fetch the list of nodes from the cluster and make it reconfigure itself (see also here). That’s finally possible because MySQL Group Replication shows the list of nodes in a […]



Comment on MySQL 5.7 – HTTP Plugin for MySQL by HTTP Plugin for MySQL | Joseph Scott

Mon, 06 Oct 2014 16:24:23 +0000

[…] There is a HTTP plugin for MySQL available at labs.mysql.com. Ulf Wendel covers it in more detail. […]



Comment on Is the HTTP Plugin for MySQL secure? by Wlad

Thu, 02 Oct 2014 21:06:32 +0000

Ulf, how do you cannot encode binary data with unicode characters, can you expand on this a bit? I thought it would be impossible and you would have to encode hex or base64 or something like this. (I also fail to understand why are you using \uXXXX for German , if you have UTF8, but this does not seem to be illegal, just odd)



Comment on MySQL 5.7 – HTTP Plugin for MySQL by john

Wed, 01 Oct 2014 06:38:03 +0000

"Restrict the user to whateer you want." - yes, this is the only way, BUT this is not enough to avoid for example list out from a table payment_transaction user_id = Y meanwhile the user is X. -> you allowed sql user to list table payment_transaction So the PHP was a wrapper to handle it with log in sessions etc.



Comment on MySQL 5.7 – HTTP Plugin for MySQL by admin

Mon, 29 Sep 2014 16:44:33 +0000

I'm taking a rest today. Let me give a more complete view in an extra blog posting, deal?



Comment on MySQL 5.7 – HTTP Plugin for MySQL by admin

Mon, 29 Sep 2014 14:51:24 +0000

I mean what I say: the SQL endpoint runs any SQL you want. It runs SQL as a certain MySQL user. Restrict the user to whateer you want.



Comment on MySQL 5.7 – HTTP Plugin for MySQL by illu

Mon, 29 Sep 2014 14:44:43 +0000

I think you mean to say: "There is nothing built in to resolve sql injections or unwanted bad sql queries"



Comment on MySQL 5.7 – HTTP Plugin for MySQL by admin

Mon, 29 Sep 2014 12:20:05 +0000

The SQL endpoint runs any SQL you want and allow for the MySQL user used to execute it.



Comment on MySQL 5.7 – HTTP Plugin for MySQL by john

Mon, 29 Sep 2014 12:14:45 +0000

How can this plugin resolve the sql injections and the unwanted bad sql-s?



Comment on MySQL 5.7 – HTTP Plugin for MySQL by MySQL 5.7 – HTTP Plugin for MySQL | InsideMySQL

Sun, 28 Sep 2014 05:01:42 +0000

[…] post MySQL 5.7 – HTTP Plugin for MySQL appeared first on Ulf […]



Comment on PECL/mysqlnd_ms: Distributed Transaction/XA support coming, e.g. for MySQL Fabric? by PECL/mysqlnd_ms: summer time, (connection) pool time - Fabric support internals | Ulf WendelUlf Wendel

Thu, 07 Aug 2014 21:36:35 +0000

[…] but not least, there is our new distriuted/XA transaction support. Imagine you need a transaction that spawns two shards. You start it on one shard and […]



Comment on PoC: Using a Group Communication System (Isis2) to improve MySQL Replication HA by PECL/mysqlnd_ms: summer time, (connection) pool time - Fabric support internals | Ulf WendelUlf Wendel

Thu, 07 Aug 2014 21:26:50 +0000

[…] any of the better NoSQL clustering solutions. Automatic master failover? Considered a user task. Have clients discover the new master automatically? Unthinkable! Sharding for MySQL Replication? Considered a user task. MySQL Proxy? Still alpha. […]



Comment on PECL/mysqlnd_ms: Distributed Transaction/XA support coming, e.g. for MySQL Fabric? by PECL/mysqlnd_ms: how failed XA transactions get fixed | Ulf WendelUlf Wendel

Tue, 29 Jul 2014 10:25:03 +0000

[...] transactions use the two-phase commit protocol, which is a blocking protocol. Please, see also my previous blog post on the nature of the protocol and MySQL implementation limitation. If the client that drives the XA [...]



Comment on Contact by bury

Mon, 14 Jul 2014 05:41:51 +0000

HI dear: When I used mysqlnd_ms (1.6.0-alpha (10600)) for mysql fabric in php. my mysqlnd_ms.config_file's json is ----------------------------------------------------- { "myapp":{ "fabric":{ "hosts": [ { "host": "192.168.1.192", "port":"32274"} ] } } } --------------------------------- but when i the php have a error . i use gdb to see the Coredump it's like this --------------------------------------------------------------------------------------------------------------- Reading symbols from /usr/local/php/lib/php/extensions/no-debug-non-zts-20121212/memcache.so...done. Loaded symbols for /usr/local/php/lib/php/extensions/no-debug-non-zts-20121212/memcache.so Reading symbols from /usr/local/php/lib/php/extensions/no-debug-non-zts-20121212/mysqlnd_ms.so...done. Loaded symbols for /usr/local/php/lib/php/extensions/no-debug-non-zts-20121212/mysqlnd_ms.so Core was generated by `php fabric.php'. Program terminated with signal 11, Segmentation fault. #0 0x000000000088f035 in php_mysqlnd_conn_data_get_host_info_pub () Missing separate debuginfos, use: debuginfo-install cyrus-sasl-lib-2.1.23-13.el6_3.1.x86_64 freetype-2.3.11-14.el6_3.1.x86_64 glibc-2.12-1.132.el6_5.2.x86_64 keyutils-libs-1.4-4.el6.x86_64 krb5-libs-1.10.3-15.el6_5.1.x86_64 libcom_err-1.41.12-18.el6.x86_64 libcurl-7.19.7-37.el6_5.3.x86_64 libgcc-4.4.7-4.el6.x86_64 libidn-1.18-2.el6.x86_64 libjpeg-turbo-1.2.1-3.el6_5.x86_64 libpng-1.2.49-1.el6_2.x86_64 libselinux-2.0.94-5.3.el6_4.1.x86_64 libssh2-1.4.2-1.el6.x86_64 libstdc++-4.4.7-4.el6.x86_64 libxml2-2.7.6-14.el6_5.2.x86_64 nspr-4.9-1.el6.x86_64 nss-3.13.3-6.el6.x86_64 nss-softokn-freebl-3.12.9-11.el6.x86_64 nss-util-3.13.3-2.el6.x86_64 openldap-2.4.23-26.el6.x86_64 openssl-1.0.1e-16.el6_5.14.x86_64 zlib-1.2.3-29.el6.x86_64 (gdb) -------------------------------------------------------------------------------------------------------------------- how can i do ?



Comment on PECL/mysqlnd_ms: Distributed Transaction/XA support coming, e.g. for MySQL Fabric? by admin

Sun, 06 Jul 2014 21:45:52 +0000

Hi Joarder, Fabric is at very early development stage after all. The Fabric development team is very much aware of the problems that arise when sharding is used. My take is that we simply have not learned yet what users wants. The initial assumption is that (power) users are very much aware of the drawbacks. The power users we spoke to initally are fine with the limits. That said, there are plans to handle very basic cross-shard queries in the drivers, stuff like basic join operations. This would make sharding a bit more transparent. Of course, there are other options but which one is best? Best as in which one do our users prefer, not which is best from an academic standpoint. XA/2PC is one tiny piece in the puzzle. What I'm doing is hardly more than fishing for input... I do not know whether this piece will be of relevance for Fabric in the future - no decision made yet.



Comment on PECL/mysqlnd_ms: Distributed Transaction/XA support coming, e.g. for MySQL Fabric? by Joarder Kamal

Sun, 06 Jul 2014 04:44:13 +0000

Very interesting posting. I was searching this exact contents on how Fabric manages XA transactions and came across this page :) This was the missing part of the puzzle I was looking for. I'm working on a research project and currently exploring options to develop a middleware to handle distributed transactions for sharded MySQL clusters. I guess mysqlnd_ms might help me upto a certain level in this case. Recently I also asked a question about Fabric in the MySQL forum and you might be also interested to have a look http://forums.mysql.com/read.php?144,616918,616918#msg-616918. Thanks again.



Comment on MySQL Fabric’s initial decision on data models and transactions put into perspective by PECL/mysqlnd_ms: Distributed Transaction/XA support coming, e.g. for MySQL Fabric? | Ulf WendelUlf Wendel

Fri, 13 Jun 2014 16:02:06 +0000

[...] transactions… Lets keep the focus narrow and talk XA only, let’s forget about cross shard joins here. Even if the PHP MySQL APIs had proper XA support (and, of course, it will be so in no more than [...]



Comment on Connecting MySQL and the Dojo Toolkit (JavaScript): basics by frank

Sat, 31 May 2014 19:28:24 +0000

hi Admin, I am really happy to have found your blog on the dojo with mysql. It is just great. Also the way you explained jsonP was awesome. After reading several sites on jsonp, this is the first blog that I did not have to read the blog twice. Waiting for your next blog on Dojo.. cheers!!!! Frank



Comment on MySQL 5.7: SQL functions for JSON by Leonardo Eloy

Fri, 30 May 2014 19:10:02 +0000

Hi Sveta, Is there any plan to upgrade this to production use? Thanks!



Comment on Train your elePHPants for the PHP Unconf EU – Palma de Majorca! by PHP Unconference Europe, a legend. But lacking PHP... | Ulf WendelUlf Wendel

Wed, 21 May 2014 17:30:56 +0000

[...] own elePHPants, well prepared and trained for the trip, became friends with him. However, I think, the big one is not quite such a sports gun as the [...]






Comment on PHP Summit presentation: NoSQL in MySQL by Late Today, Early Tomorrow » Blog Archive PostgresSQL 9.4 adresses NoSQL developers » Late Today, Early Tomorrow

Tue, 20 May 2014 09:36:48 +0000

[...] This allows fast key-value retrieve and store to InnoDB tables. Take a look at this very nice presentation of Ulf Wendel from Oracle how this is intended to work. It was a first major attempt of [...]



Comment on JS Unconf Hamburg 2014 by Train your elePHPants for the PHP Unconf EU - Palma de Majorca! | Ulf WendelUlf Wendel

Thu, 01 May 2014 17:11:39 +0000

[...] flights from Paris/London/Berlin to PMI are still available starting at ~120 Euros. During JS Unconf, I’ve heard of people sharing holiday homes (including Wifi ) for very low rates. If you [...]



Comment on JS Unconf Hamburg 2014 by JS Unconference EU im Geomatikum in Hamburg | Die wunderbare Welt von Isotopp

Tue, 29 Apr 2014 08:11:43 +0000

[...] Bericht und Bilder: http://blog.ulf-wendel.de/2014/js-unconf-hamburg-2014/ [...]



Comment on PHP mysqlnd memory optimizations: from 49MB to 2MB by hyh

Sat, 26 Apr 2014 17:34:43 +0000

Do you know when the solution may be compatible with PDO?



Comment on PHP mysqlnd memory optimizations: from 49MB to 2MB by Link del giorno 11/04/2014 | Catnic Blog

Fri, 11 Apr 2014 07:14:22 +0000

[...] Ottimizzare la memoria per mysqlnd [...]



Comment on The performance penalty of the early MySQL Fabric support for PHP by Ulf Wendel: The performance penalty of the early MySQL Fabric support for PHP | facebooklikes

Thu, 13 Mar 2014 17:23:40 +0000

[...] his latest post Ulf Wendel looks at the performance issues around the recently introduced MySQL Fabric support [...]



Comment on Using MySQL Fabric from any programming language by The performance penalty of the early MySQL Fabric support for PHP – Ulf Wendel | WarWebDev

Thu, 13 Mar 2014 08:30:00 +0000

[...] exactly a “Fabric aware” driver or application talks to Fabric is implementation dependent. Figures given for PHP must [...]



Comment on Using MySQL Fabric from any programming language by The performance penalty of the early MySQL Fabric support for PHP | Ulf WendelUlf Wendel

Wed, 12 Mar 2014 22:42:34 +0000

[...] exactly a “Fabric aware” driver or application talks to Fabric is implementation dependent. Figures given for PHP [...]



Comment on PDO_MYSQLND: R[a|u]mbling and a breeze of progress by HDBC-MySQLのRow retrieval was canceled by mysql_stmt_close() call対策 – 放置演算子

Wed, 12 Mar 2014 16:23:26 +0000

[...] PDO_MYSQLND: R[a|u]mbling and a breeze of progress | Ulf WendelUlf Wendel [...]



Comment on Using MySQL Fabric from any programming language by The early MySQL Fabric sharding support for PHP | Ulf WendelUlf Wendel

Fri, 07 Mar 2014 23:08:33 +0000

[...] you need a “Fabric aware driver” for this but it does not list one for PHP. First, you don’t necessarily need new drivers! Second, the "driver" for PHP is the current development version of the PECL/mysqlnd_ms [...]



Comment on Using MySQL Fabric from any programming language by Using MySQL Fabric from any programming language – Ulf Wendel | WarWebDev

Wed, 05 Mar 2014 01:05:17 +0000

[...] by Planet PHP, read more at the original (another 20153 [...]



Comment on PECL/mysqlnd_ms 1.5 + PHP 5.5.0 + MySQL 5.6 = better load balancing by micu

Tue, 11 Feb 2014 07:55:54 +0000

Does anyone know if PDO_MySQL has been modified in php 5.5 to support the new mysqlnd API calls? Thanks



Comment on PHP Memcache access to MySQL 5.7, faster? Redis? by mindy

Tue, 31 Dec 2013 14:09:48 +0000

great, do you test redis by phpredis?



Comment on PHP Memcache access to MySQL 5.7, faster? Redis? by Ulf Wendel: PHP Memcache access to MySQL 5.7, faster? Redis? | htaccess

Fri, 13 Dec 2013 19:17:26 +0000

[...] his site Ulf Wendel shows an alternative use for the PHP Memcache functions – using them to query MySQL tables (InnoDB) in much the same way. He also tosses in Redis as another version to compare the [...]



Comment on PHP Memcache access to MySQL 5.7, faster? Redis? by Ulf Wendel: PHP Memcache access to MySQL 5.7, faster? Redis? | facebooklikes

Fri, 13 Dec 2013 18:51:28 +0000

[...] his site Ulf Wendel shows an alternative use for the PHP Memcache functions – using them to query MySQL tables (InnoDB) in much the same way. He also tosses in Redis as another version to compare the [...]



Comment on PHP Memcache access to MySQL 5.7, faster? Redis? by Ulf Wendel

Fri, 13 Dec 2013 11:02:47 +0000

The past five years have been busy times for MySQL. Of course, NoSQL is a threat and it may have impacted our way somewhat, albeit this particular story began with Handlersocket and AFAIK not as a NoSQL reply. Whatever, NoSQL is not only about performance and crazy scalability. Some systems are small, have such neat features and are that easy to use - as MySQL was 15 years ago? Maybe it is impossible to hide the complexity if you want the fastest monster. Still, cleanup, some neat new features for developers... how about, for example, this feature request from 2005: http://bugs.mysql.com/bug.php?id=13830 , "Support for ROW, ARRAY, and MULTISET type constructors" - the guy that filed this bug is pretty well known in the PHP world: Sebastian Bergmann.



Comment on PHP Memcache access to MySQL 5.7, faster? Redis? by Peter Laursen

Thu, 12 Dec 2013 21:12:21 +0000

I agree that performance on high-end systems has been *over-focused* for almost 5 years compared to ease of use for the average (and average+) user. Almost every new MySQL release adds complexity and there is practically never any cleanup in old oddities.



Comment on 1.5x … 9x faster queries with PHP and MySQL 5.6, really? by PHP Memcache access to MySQL 5.7, faster? Redis? | Ulf WendelUlf Wendel

Thu, 12 Dec 2013 18:17:32 +0000

[...] meetup visitor roll his eyes and say “yeah, Oracle, *yawn*…”. I’ve repeated my plain PHP benchmarks on an i3 desktop. And, I’ve added Redis to the [...]



Comment on Not only SQL injection: I don’t trust you! by Raghu Veer

Mon, 09 Dec 2013 07:17:20 +0000

I regularly use filter to (sanitize input first and then validate, if input is url, int, float, email etc) and then use PDO Prepared Statements when inserting the input into the MySQL Database. While I am interested to see your slides, I find 404 in mysql website for the webinar link. Please provide the link to webinar if it is hosted elsewhere, thank you



Comment on MySQL 5.7 Fabric: any good? by MySQL Fabric with MariaDB Galera Cluster ? « Serge Frezefond 's blog

Thu, 31 Oct 2013 15:46:54 +0000

[...] VN (Narayanan Venkateswaran) Installing MySQL Fabric on Windows  2013-10-03 Todd Farmer MySQL 5.7 Fabric: any good?  2013-09-23 Ulf [...]



Comment on MySQL Fabric’s initial decision on data models and transactions put into perspective by MySQL Fabric with MariaDB Galera Cluster ? « Serge Frezefond 's blog

Thu, 31 Oct 2013 15:35:14 +0000

[...]    InfiniDB column store moves to open source ! Congrats ! » MySQL Fabric with MariaDB Galera Cluster ? MySQL Fabric is a very promising sharding framework. If I take Ulf Wendel definition of MySQL Fabric : [...]



Comment on MySQL 5.7: SQL functions for JSON by MySQL Fabric's initial decision on data models and transactions put into perspective | Ulf WendelUlf Wendel

Wed, 16 Oct 2013 23:50:18 +0000

[...] document = shard + relevant global tables + x. All MySQL has to offer here are some, limited JSON functions. On an aside: this is only one of many ways to argue why MySQL should have strong JSON/BSON [...]






Comment on MySQL 5.7: SQL functions for JSON by MySQL soll jetzt auch JSON bekommen | Klaus Ahrens: News, Tipps, Tricks und Fotos

Thu, 10 Oct 2013 15:57:07 +0000

[...] Blog-Beitrag von Ulf Wendel bestätigt jetzt, dass daran gearbeitet wird. Wendel beschreibt ausführlich die im [...]



Comment on MySQL 5.7: SQL functions for JSON by Sveta Smirnova

Wed, 09 Oct 2013 22:54:58 +0000

Hi Ulf, thank you very much for the deep analysis! I really appreciate it. I created few bug reports from cases which you found. There also were bug reports, reported earlier. Please subscribe to them to receive updates. > With my source build the resulting .so file got a name – libmy_json_udf.so – different from the one – libmy_json.so – assumed in the README. We already had http://bugs.mysql.com/bug.php?id=70392 about the same issue. > Regarding JSON_valid. > "a", [1], ["a"], [1,{"a":"bc"}] This is by design: I implemented definition of the object from json.org, see diagram on the right side. But looks I was too strict. But according to http://www.ietf.org/rfc/rfc4627.txt?number=4627 I only should reject "a". I reported the bug: http://bugs.mysql.com/bug.php?id=70567 {"a":false,} {"a":"\z"} - both cases reported as http://bugs.mysql.com/bug.php?id=70568 {"a":nuLL} {"a":True} {"a":FaLSE} - Oh, I see the issue of compatibility with http://www.ietf.org/rfc/rfc4627.txt?number=4627, three cases reported as http://bugs.mysql.com/bug.php?id=70569 {"a":"\u01"} - http://bugs.mysql.com/bug.php?id=70570 > Keys and arrays work as expected. Array keys are zero-based. The first element in an array is accessed through the key “0″, the second element through key “1″ and so forth. Note, however, that you have to use strings. There is a bug report for this already: http://bugs.mysql.com/bug.php?id=70393 > If one of following documents does not contain an opening curly bracket first documents merged are returned and warning is generated. Oh, it was my original wish, but I stacked with the fact that UDFs do not allow to produce any warning for user errors after you finish job in *_init. Unfortunately, in this case file README should be updated: http://bugs.mysql.com/bug.php?id=70573 > And the last sentence from the README is: [returns] NULL if first document does not contain an opening curly bracket. This is pure bug: http://bugs.mysql.com/bug.php?id=70574. > Invalid JSON returned: > > mysql> select json_replace('{"a":[1]}', "a", "b")\G This is pure bug: http:/[...]



Comment on MySQL 5.7 Fabric: any good? by A new kid in the MySQL sharding world « Serge Frezefond 's blog

Wed, 09 Oct 2013 18:58:17 +0000

[...] VN (Narayanan Venkateswaran) Installing MySQL Fabric on Windows  2013-10-03 Todd Farmer MySQL 5.7 Fabric: any good?  2013-09-23 Ulf Wendel — Writing a Fault-tolerant Database Application using MySQL [...]



Comment on MySQL 5.7: SQL functions for JSON by Mark Callaghan

Wed, 09 Oct 2013 00:29:53 +0000

Excellent post, but InnoDB can/will store more than the first 768 bytes of a blob column inline in many cases.



Comment on MySQL 5.7: SQL functions for JSON by Ammon Sutherland

Wed, 09 Oct 2013 00:16:39 +0000

Glad to see someone else found these on the labs site. Your writeup was good and thorough and helped explain some of the issues I found with the key-value particulars when testing briefly. Thanks.



Comment on Searching document stores in 2013: from 1983 to SQL:2003 in a blink? by MySQL 5.7: SQL functions for JSON overview | Ulf WendelUlf Wendel

Tue, 08 Oct 2013 22:44:37 +0000

[...] It may sound a bit though but this is a case where you see the difference between a MeeTooSQL and a system that has had the time and resources to add syntactic sugar. With a bit of syntactic sugar, say “SELECT document.member[offset]” this could look much more appealing. On the contrary, I am talking syntaxtic sugar only! Syntactic sugar is really hard to add with todays MySQL. The MySQL parser is not modular, and at the hearth of the entire system, which forbids icing on the cake in the making. Not to speak of ignoring the SQL standard and the option to support such a notation in UDFs at the cost of parsing string arguments (over processing a **va_args list). Still, as a developer, … See also, Searching document stores in 2013: from 1983 to SQL:2003 in a blink? [...]



Comment on MySQL 5.7 Fabric: any good? by Mats Kindahl

Fri, 27 Sep 2013 05:17:36 +0000

Hi Justin, Regarding the use of "sharded tables" and "unsharded tables" it seems to be more common to call the unsharded tables "global". Personally, I have no strong opinion, so either one is fine and will probably vary depending on the context where one describes it. In Fabric, is architected to support multiple mappings/indexes (I assume that is what you call a virtual schema) in the same manner: by restricting it to use a single key per mapping. It is possible to shard tables using multiple keys, but that requires executing a scatter-gather query whenever a shard is accessed, which affects performance and is also more complicated to implement. For the first version, we have just assumed there is a single mapping because the handling of global updates were not entirely sorted out. MySQL Fabric uses a global server/group to execute global updates and then use replication to replicate the updates to the shards. If you note how the computation of the shard is done, MySQL Fabric accept a list of tables for the transaction and use that to automatically compute which shard mapping it belongs to and can thereafter compute the shard from the shard key. The case that were not fully sorted out is what to do when all tables where unsharded/global. Currently, a global server/group was assigned per shard mapping, but going forward we intend to have a global server/group for each Fabric instance instead.






Comment on MySQL 5.7 Fabric: any good? by admin

Mon, 23 Sep 2013 17:25:12 +0000

It is a good idea to speak of sharded and unsharded tables. I should follow your wording. Currently, Fabric does not prevent anything. Shards even accept writes for keys for which they are not responsible. This is something future versions should not allow, as far as I know. But, you are raising another issues that - to my knowledge - is not covered yet. Nobody thought about it. The few existing fabric aware drivers don't even cover all basics yet :-/. I guess, we can learn alot from http://code.google.com/p/shard-query/



Comment on MySQL 5.7 Fabric: any good? by Justin Swanhart

Mon, 23 Sep 2013 17:09:44 +0000

I prefer the term "sharded tables" and "unsharded tables" . Unsharded tables are duplicated on all shards, while sharded tables are distributed between shards based on the shard key. Also, how does fabric prevent the user from joining between tables that are sharded by different keys? Shard-Query prevents this by only allowing one shard key per virtual schema.



Comment on MySQL 5.7 Fabric: any good? by PHP Unconference Hamburg 2013: hungry for more | Ulf WendelUlf Wendel

Mon, 23 Sep 2013 15:05:32 +0000

[...] 7th floor and, well, business as usual: 200+ visitors, high quality sessions, especially my own on MySQL 5.7 with built-in sharding , 40 Euros entry fee for what costs 400+ Euros [...]



Comment on PoC: Using a Group Communication System (Isis2) to improve MySQL Replication HA by MySQL 5.7 Fabric: any good? | Ulf WendelUlf Wendel

Mon, 23 Sep 2013 10:22:21 +0000

[...] failure detector, activate. E.g., use built-in (imperfect) failure [...]



Comment on PoC: Using a Group Communication System (Isis2) to improve MySQL Replication HA by Baron Schwartz

Thu, 08 Aug 2013 19:11:49 +0000

You said, If the probability of one system to fail is 1:1000, then the probability of two systems failing at the same time is 1:10000. This is true, if failures are independent events. The problem is that most of the time, they aren't, in systems like this. That's why distributed systems seem to fail a lot more than we think they should.



Comment on Connecting MySQL and the Dojo Toolkit (JavaScript): basics by Connecting MySQL and the Dojo Toolkit: MySQL data store | Ulf WendelUlf Wendel

Mon, 24 Jun 2013 18:12:55 +0000

[...] it is to populate a sortable, … data grid using a store. Based on the proxy sketched in the previous post, I’ve prototyped two MySQL stores for dojo. One maps to a MySQL table, the other one maps to [...]



Comment on Supercharging PHP MySQL applications using the best API by Henry McDaniel

Sat, 22 Jun 2013 19:21:05 +0000

People who know how to write good code already had safe code with the old mysql functions. Maybe the real problem is that people who don't know how to code kept copy and pasting other people's shoddy code.






Comment on Supercharging PHP MySQL applications using the best API by Frage zur Umstellung mysql -> mysqli - php.de

Thu, 28 Mar 2013 19:32:07 +0000

[...] doch. Und wass ist denn anders bei mysqli, außer dem i hinten drann? Schau dir z.B. mal diesen Blog-Artikel von Ulf Wendel (*) an. * Ulf Wendel ist Entwickler des MySQL Native Drivers, der in PHP für die 3 [...]



Comment on Searching document stores in 2013: from 1983 to SQL:2003 in a blink? by Open Sourcing Akiban | The Akiban Blog - Database Flexibility without Compromise

Fri, 22 Mar 2013 22:56:27 +0000

[...] using Lucene) * entity (or group) indexes * REST access * nested SQL (I noticed Ulf Wendel recently discussed something like [...]



Comment on PECL/mysqlnd_ms 1.5 + PHP 5.5.0 + MySQL 5.6 = better load balancing by admin

Wed, 20 Mar 2013 19:07:31 +0000

Well, how about "Download Latest" to get the latest unreleased development version...



Comment on PECL/mysqlnd_ms 1.5 + PHP 5.5.0 + MySQL 5.6 = better load balancing by PHP: PECL/mysqlnd_ms 1.5 alpha, mysqlnd_qc 1.1 alpha releases | Ulf WendelUlf Wendel

Wed, 20 Mar 2013 18:57:00 +0000

[...] and the Query Cache plugin (PECL/mysqlnd_qc) for PHP mysqlnd. PECL/mysqlnd_ms 1.5 is a step up from the previous version whereas PECL/mysqlnd_qc 1.1 is nothing but a maintenance [...]



Comment on PECL/mysqlnd_ms 1.5 + PHP 5.5.0 + MySQL 5.6 = better load balancing by b0Rman

Fri, 15 Mar 2013 18:48:07 +0000

Nope, there is no v1.5 yet : http://pecl.php.net/package/mysqlnd_ms



Comment on PECL/mysqlnd_ms 1.5 + PHP 5.5.0 + MySQL 5.6 = better load balancing by admin

Thu, 14 Mar 2013 15:04:34 +0000

As the name PECL/mysqlnd_ms says: PECL (pecl.php.net) ....



Comment on PECL/mysqlnd_ms 1.5 + PHP 5.5.0 + MySQL 5.6 = better load balancing by b0Rman

Thu, 14 Mar 2013 05:23:34 +0000

Is it possible to get 1.5 dev.version somewhere? :)



Comment on PECL/mysqlnd_ms 1.5 + PHP 5.5.0 + MySQL 5.6 = better load balancing by admin

Wed, 13 Mar 2013 17:01:06 +0000

I probably should have stressed out that I'm talkig about the development version (PECL/mysqlnd_ms 1.5) - there is no release yet. However, note that the improved loadbalancing requires PHP 5.5.0. PHP 5.5 is still alpha.



Comment on PECL/mysqlnd_ms 1.5 + PHP 5.5.0 + MySQL 5.6 = better load balancing by b0Rman

Wed, 13 Mar 2013 12:01:11 +0000

That's nice, but where can we get PECL/mysqlnd_ms 1.5? It's still under construction and there is no even alpha of 1.5 :(



Comment on PECL/mysqlnd_ms 1.5 + PHP 5.5.0 + MySQL 5.6 = better load balancing by PECL/mysqlnd_ms 1.5 + PHP 5.5.0 + MySQL 5.6 = better load balancing | 傻子日志

Sun, 24 Feb 2013 23:55:56 +0000

[...] post PECL/mysqlnd_ms 1.5 + PHP 5.5.0 + MySQL 5.6 = better load balancing appeared first on Ulf [...]



Comment on 1.5x … 9x faster queries with PHP and MySQL 5.6, really? by admin

Thu, 31 Jan 2013 11:53:32 +0000

That is a valid question. There are three challenges: a) one would have to change the protocol, which is kind of a holy cow b) the protocol handling has to be modular inside the server, which is not the case c) you would likely have to create a new connect without auth/perm overhead d) extend all drivers - including 3rd party ones It's way easier to wrap a new protocol around an internal data access API (here: InnoDB [embedded] API) than it is to extend the existing protocol code. MySQL was never designed to have pluggable or modular or easily extensible protocol. Refactoring will take time. More time than opening an additional port. Plus, Memcache integration/compatibility was a major goal.



Comment on 1.5x … 9x faster queries with PHP and MySQL 5.6, really? by Marc Bennewitz

Thu, 31 Jan 2013 11:35:37 +0000

Thanks for this great article! What I was question to me: Wouldn't it be a better strategy to implement such a simple KV protocol directly within the MySQL protocol? Than the connection overhead goes away.



Comment on Searching document stores in 2013: from 1983 to SQL:2003 in a blink? by admin

Fri, 18 Jan 2013 15:38:07 +0000

Thanks for taking the time to correct me on "scalar" vs. atomic! I was wondering whether one could store JSON documents in a somewhat reasonal way inside a RDBMS.



Comment on Searching document stores in 2013: from 1983 to SQL:2003 in a blink? by Fernando

Fri, 18 Jan 2013 15:23:18 +0000

I think you're confusing the relational model with relational-inspired databases when you explain 1NF. My understanding of theory is that every relation is in 1NF by definition. Atomic values don't have to be scalar, they just have to be atomic on their domain. You can see this in action even in MySQL, as the spatial data types aren't scalar, yet they are atomic. So what I understand from 1NF is that, basically, 'every attribute of a relation has to have a type'. Atomicity here is defined only in relation to this type. It is true, however, that people like Date discourage the use of 'relation valued attributes' from a design point of view ("... let me just say that relvars with RVAs tend to look very much like the hierarchic structures found in older, nonrelational systems like IMS ..." quoting from "SQL and Relation Theory"), but those would be in 1NF any way.



Comment on Searching document stores in 2013: from 1983 to SQL:2003 in a blink? by admin

Fri, 18 Jan 2013 08:15:27 +0000

XML has a schema, JSON has not. Why use XPath, if there's SQL for nested data...



Comment on Searching document stores in 2013: from 1983 to SQL:2003 in a blink? by gggeek

Thu, 17 Jan 2013 23:03:31 +0000

You should investigate into the support for xml storage of existing databases. They probably already do allow something close to your goals, using xpath for queries...



Comment on Searching data in NotOnlyMySQL databases: a rich query language? by Searching document stores in 2013: from 1983 to SQL:2003 in a blink? | Ulf WendelUlf Wendel

Thu, 17 Jan 2013 16:39:55 +0000

[...] to are rich and powerful, has to be checked on a case by case basis. For example, I have my doubts (english version, german version) about the MongoDB query language. MapReduce is certainly great for distributed [...]



Comment on … and a Happy New Year! by … and a Happy New Year! | codegooroo

Sun, 30 Dec 2012 21:35:12 +0000

[...] post … and a Happy New Year! appeared first on Ulf [...]






Comment on Vote NO for MySQL – Election 2012: NoSQL by Vote NO for MySQL – Election 2012: NoSQL | codegooroo

Wed, 12 Dec 2012 05:10:36 +0000

[...] post Vote NO for MySQL – Election 2012: NoSQL appeared first on Ulf [...]



Comment on Supercharging PHP MySQL applications using the best API by A Smattering of Selenium #132 « Official Selenium Blog

Thu, 06 Dec 2012 01:56:47 +0000

[...] Supercharging PHP MySQL applications using the best API is useful when testing your app in general, but also remember that if you are trusting what the browser is telling you and not verifying it with the database you are asking for a world of hurt Share this:EmailFacebookLinkedInDiggTwitterGoogle +1Like this:LikeBe the first to like this. Leave a Comment [...]



Comment on Supercharging PHP MySQL applications using the best API by PHP Digest: Simplified Password Hashing, Creation of PHP Framework Using Symfony2 Components, Xdebug Tutorial and Much More | Zfort Group Blog

Tue, 04 Dec 2012 10:17:51 +0000

[...] Supercharging PHP MySQL applications using the best API – at the moment there are three types of PHP interface to work with MySQL: ext / mysql, ext / mysqli and PDO_MySQL. In php.internals the discussion has occurred whether to remove the obsolete ext / mysql from the delivery set by default and transfer it to PECL. The author uncovers the topic in detail, providing a comprehensive comparison of three interfaces mentioned above. [...]