tag:blogger.com,1999:blog-5552895785228669482.post5023275149829499623..comments2024-03-02T07:39:03.908+02:00Comments on Monty says: Oops, we did it again (MySQL 5.1 released as GA with crashing bugs)Montyhttp://www.blogger.com/profile/06049512911785594864noreply@blogger.comBlogger34125tag:blogger.com,1999:blog-5552895785228669482.post-50761198823999221262009-03-25T03:37:00.000+02:002009-03-25T03:37:00.000+02:00Monty, thank you for the post. I just started work...Monty, thank you for the post. I just started working a sysadmin at site with Mysql database, and had several days nightmare - segfault. <BR/><BR/>When we traced down it to a last stored procedure call we was ready to give up. Simple insert to ndm table crashed mysqld (both 5.1.25rc and 5.1.32). Recreating table did not help. But then my developer recreated index and... it's working now.Shebnikhttps://www.blogger.com/profile/07920397085996163547noreply@blogger.comtag:blogger.com,1999:blog-5552895785228669482.post-1785130637595178662009-02-18T21:16:00.000+02:002009-02-18T21:16:00.000+02:00MySQL is a class of extremely critical software th...MySQL is a class of extremely critical software that people absolutely depend upon. That means that they should be doing daily builds and automated bug checking. I've read <A HREF="http://www.joelonsoftware.com" REL="nofollow">Joel Spolsky</A>, and according to his metrics, this is one of the ten most important things a company can do to improve the quality of the software it releases. With SUN having bought this company it would presume they have the resources to do this. Hell, I am a one-man software company and when I was working on a project I did regular daily builds and (while not as rigorous as a company that can afford automated testing software) I still went through and ran tests on everything I changed, <I>on a daily basis</I> to make sure I didn't break anything.<BR><BR/>Paul Robinson - <A HREF="http://paul-robinson.us" REL="nofollow">My Blog</A><BR><BR/>"The lessons of history teach us - if they teach us anything - that no one learns the lessons that history teaches us."Paul Robinsonhttps://www.blogger.com/profile/08691621235704322096noreply@blogger.comtag:blogger.com,1999:blog-5552895785228669482.post-64271592808696930762009-02-09T15:43:00.000+02:002009-02-09T15:43:00.000+02:00We have been using MySQL 5.1.22 on an cluster for ...We have been using MySQL 5.1.22 on an cluster for 14 months now, with very few problems, we had some issues with the memory management if you have very big databases, but it was mainly due to the lack of explanation on the settings. Until I have a major crash on my hand or a need to reinstall, I'm planning to continue using the community version.<BR/><BR/>Thanks for the post though, Some of my collaborators were asking for the updating after the announcement, not going to happen now that I've read thisUnknownhttps://www.blogger.com/profile/10819572516096467725noreply@blogger.comtag:blogger.com,1999:blog-5552895785228669482.post-69285615989922410182009-02-06T18:42:00.000+02:002009-02-06T18:42:00.000+02:00I think most people in the closed-source database ...I think most people in the closed-source database world recognise the release inflation that has taken place over the years.<BR/><BR/>GA == beta<BR/>RC == alpha<BR/><BR/>If, as Big Corporate(tm)'s CTO, you don't know that, you're likely going to be paying through the nose for support contracts and collecting unemployment not long after. Welcome to Oracle and DB2's business model.<BR/><BR/>For those of us that are just trying to get stuff done, if 5.1 works for us, great, if not, we roll back. Life will go on, no one expects complex software to be completely bug-freeAlex Tomichttps://www.blogger.com/profile/12973523024418202623noreply@blogger.comtag:blogger.com,1999:blog-5552895785228669482.post-59476911353698699182009-02-06T06:45:00.000+02:002009-02-06T06:45:00.000+02:00I've been a big fan of MySQL since v4. That said I...I've been a big fan of MySQL since v4. That said I think the worst thing that could have happened to MySQL was selling it to Sun. <BR/><BR/>You guys were on the right track, give the software away, charge for the support. <BR/><BR/>But once you sold to Sun it became create a marketable product on an obscenely bad timetable. <BR/><BR/>Best of luck with the new company.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-5552895785228669482.post-62217678085581144502009-01-14T03:35:00.000+02:002009-01-14T03:35:00.000+02:00cautious about MySQL 5.1 is that there are still m...cautious about MySQL 5.1 is that there are still many known and unknown fatal bugs in the new features that are still not addressed. Great post thanksInventory management softwarehttps://www.blogger.com/profile/10259038069827718281noreply@blogger.comtag:blogger.com,1999:blog-5552895785228669482.post-87632234571325135252008-12-18T19:00:00.000+02:002008-12-18T19:00:00.000+02:00I develope Complex Web Server (http://ponkrac.net/...I develope Complex Web Server (http://ponkrac.net/complex-web-server/en) as a reliable, user friendly web server with many features. MySQL server is part of content of the one.<BR/><BR/>Reliability is very high priority for me. When I downloaded mysql 5.1.30 package for Windows 32 bit I got very bad feel from MySQL. Some things is not completed. For example libmysql.dll in 5.1.30 package caused crashing of PHP & Apache server. I need to substitute libmysql.dll and I use version from 5.0.67 which works good.<BR/><BR/>Official MySQL 5.1 documentation in Windows help form (.chm) is damaged - some chapters is entirely missing and caused message box with error.<BR/><BR/>Miloslav PonkrácUnknownhttps://www.blogger.com/profile/14612259190037891841noreply@blogger.comtag:blogger.com,1999:blog-5552895785228669482.post-72008971848088745032008-12-17T02:41:00.000+02:002008-12-17T02:41:00.000+02:00II think that in this moment of transaction the c...II think that in this moment of transaction the comunity must promote a quality campaign<BR/>send a mail to Sun with this subject :MySql Ask Qualitypowleanhttps://www.blogger.com/profile/04268802018183688963noreply@blogger.comtag:blogger.com,1999:blog-5552895785228669482.post-11050200493890788742008-12-13T00:26:00.000+02:002008-12-13T00:26:00.000+02:00Ya, that staute gives me a GREAT idea for a variat...Ya, that staute gives me a GREAT idea for a variation on the theme ;)pcleddyhttps://www.blogger.com/profile/06050605714525151639noreply@blogger.comtag:blogger.com,1999:blog-5552895785228669482.post-17357457291898762812008-12-11T14:31:00.000+02:002008-12-11T14:31:00.000+02:00Better to know the truth than to do a marketing po...Better to know the truth than to do a marketing polish.<BR/><BR/>Thanks, Monty.Dr. Roy Schestowitzhttps://www.blogger.com/profile/04422170397340869575noreply@blogger.comtag:blogger.com,1999:blog-5552895785228669482.post-85253138352844478412008-12-10T12:08:00.000+02:002008-12-10T12:08:00.000+02:00"...not too many get serious crashes and losses da..."...not too many get serious crashes and losses data..."<BR/><BR/>Gosh, who would use a database engine where that is a prospect you can realistically look forward to?Paul Urbanhttps://www.blogger.com/profile/03579907598970908902noreply@blogger.comtag:blogger.com,1999:blog-5552895785228669482.post-19551155586989068342008-12-07T12:03:00.000+02:002008-12-07T12:03:00.000+02:00To clarify things a little:I think that MySQL 5.1 ...To clarify things a little:<BR/>I think that MySQL 5.1 is a good *recommended* release, especially now when MySQL/Sun is providing full support for it.<BR/><BR/>What I disagree with is giving MySQL 5.1 a GA status, which at least for me, implicates it's has no crashing or other serious bug that affects normal operation.<BR/><BR/>That said, work on MySQL 5.1 continues and if things goes well we reach this goal more sooner than later.Montyhttps://www.blogger.com/profile/06049512911785594864noreply@blogger.comtag:blogger.com,1999:blog-5552895785228669482.post-18022124756119197602008-12-05T23:28:00.000+02:002008-12-05T23:28:00.000+02:00On the heels of the long-awaited GA release of MyS...<A HREF="http://www.pythian.com/blogs/1413/log-buffer-126-a-carnival-of-the-vanities-for-dbas" REL="nofollow">On the heels of the long-awaited GA release of MySQL 5.1, Monty Widenius summed up his feeling on the release thus: Oops, we did it again.</A>Dave Edwardshttps://www.blogger.com/profile/01900990241184536017noreply@blogger.comtag:blogger.com,1999:blog-5552895785228669482.post-37457760964815247052008-12-02T23:32:00.000+02:002008-12-02T23:32:00.000+02:00Isn't it fun to be an armchair quarterback? :-)Wh...Isn't it fun to be an armchair quarterback? :-)<BR/><BR/>While I agree a bit with what Chris (above) said about a utopian view of bug-free complex software, when a serious or critical bug exists in software prior to a release, an evaluation must be made on how common the bug is seen and whether or not data loss is caused as a result.<BR/><BR/>There are a number of bugs in the MySQL server that continue to cause data loss and prevent effective recovery. With bugs like #949 being around for five years, I think it's time to "stop the line" as they say at Toyota and get that binlog issue fixed. dropping a table while a transaction is in-progress should cause either the table drop to fail or the transaction commit to fail. There are no two ways about this.<BR/><BR/>Marty - this is a great post on process improvements MySQL / Sun can make so users don't get surprised by nasties of code quality failures.<BR/><BR/>There is a point that a for-profit company must decide - release with bugs and get it out the door with support headaches and the hits to reputation versus the other way - the hits to reputation due to continuing to slip what is already perceived as a late release.<BR/><BR/>As many here know (I hope), development teams have three things to "play with" on the way to a release - time available to release, features included (scope) in the release, and quality of code released (bugs fixed and defects released). Chaos results when all three are negotiable during the project lifecycle. Nothing gets done on time if none are negotiable. When one (or preferably two) of those items is fixed, that makes it much easier to negotiate on what the "Done" requirements are for a release.<BR/><BR/>I personally prefer to negotiate scope and time, but never quality. I think Monty would agree. If I had to loose one of the two of those, I'd rather give up negotiation on time. Scope is relatively easy to adjust downward and lets the customer get something useful earlier. The problem is, there are times when adjusting scope downward will force me to adjust quality downward as well.K Bhttps://www.blogger.com/profile/16179434905732017538noreply@blogger.comtag:blogger.com,1999:blog-5552895785228669482.post-68975323008962307242008-12-02T01:33:00.000+02:002008-12-02T01:33:00.000+02:00I started building data warehouse solutions on 5.1...I started building data warehouse solutions on 5.1 about a year ago because I needed table partitioning for very large fact tables. I have not experienced any crashing or corruption problems, and performance has been very good.<BR/><BR/>As a former PeopleSoft (later Oracle) employee, I can tell you that it is unrealistic to expect that highly complex software will be released without known and unknown anomalies. GA != Utopian, bug-free software.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-5552895785228669482.post-7104432392156251332008-12-02T01:03:00.000+02:002008-12-02T01:03:00.000+02:00Note to Sun:Subject: Re: Sun's Management (and the...Note to Sun:<BR/><BR/>Subject: Re: Sun's Management (and the Marketing they choose)<BR/><BR/>Get in or get off the diving board, you look silly dipping your toe in from all the way up there.ceaslesshttps://www.blogger.com/profile/02653650687694184495noreply@blogger.comtag:blogger.com,1999:blog-5552895785228669482.post-49707736896855892462008-12-02T00:50:00.000+02:002008-12-02T00:50:00.000+02:00I disagree that "quality and features" should be t...I disagree that "quality and features" should be the sole focus of a release. Marketing concerns such as timeliness should play a part. However, if forced to choose between quality and features in order to meet a release deadline, features should be sacrificed - not quality.<BR/><BR/>Hold the release schedule and quality level constant, then add as many features as you can. If "as many as you can" is "not that many" then so be it.buddyglasshttps://www.blogger.com/profile/08482253774835776080noreply@blogger.comtag:blogger.com,1999:blog-5552895785228669482.post-53538412725546174312008-12-02T00:14:00.000+02:002008-12-02T00:14:00.000+02:00Interesting post, Monty. Thanks for sharing your t...Interesting post, Monty. Thanks for sharing your thoughts.<BR/><BR/>I am curious though - you said that the problems are with the MySQL management in Sun - have these issues been address to Marten and others, and has there been some indication that they will be resolved?<BR/><BR/>It sounds like the intersection between Sun and Community with QA is facing some troubles, but I would imagine these can be fairly easy to solve, under the premise MySQL are open to change.Jono Baconhttps://www.blogger.com/profile/02369045747003603551noreply@blogger.comtag:blogger.com,1999:blog-5552895785228669482.post-42348671374568604422008-12-01T23:30:00.000+02:002008-12-01T23:30:00.000+02:00Thanks for a very thorough follow-up Monty, and I'...Thanks for a very thorough follow-up Monty, and I'm tipping my hat to your constant fights for the quality of MySQL. That's the spirit that's made me believe in the product for the past 10+ years.Unknownhttps://www.blogger.com/profile/04466184994868212679noreply@blogger.comtag:blogger.com,1999:blog-5552895785228669482.post-48702084391597770382008-12-01T20:41:00.000+02:002008-12-01T20:41:00.000+02:00well 5.1 offers some great things to play with at-...well 5.1 offers some great things to play with at-least.Dathan Pattishallhttps://www.blogger.com/profile/00356367514107959723noreply@blogger.comtag:blogger.com,1999:blog-5552895785228669482.post-61878219624704677342008-12-01T13:27:00.000+02:002008-12-01T13:27:00.000+02:00Yes, Community first! Thanks for the post. It push...Yes, Community first! Thanks for the post. It pushed to write an article, I wanted to write for some time: http://www.perspektive89.com/2008/community_first<BR/><BR/>I am supporting your view and would like to extend what you are writing to free software/content/infrastructure/wireless communities in general. I discuss this in regards to Freifunk free wireless (http://freifunk.net) and LXDE (http://lxde.org), the fastest full featured netbook desktop.<BR/><BR/>I believe communities must take the lead in order to make and keep a project on the bleeding edge and offer quality code and stability. We should work together with companies and exchange resources. Both can profit. In the end open and free community projects are all about cooperation. However, companies and communities have different agendas. You will find brilliant and ¨shiny¨ people in free software communities, who just would not work in some companies and some just would not fit in. Communities can integrate many different people. The motivation of these people goes beyond monetary interests. That is why the free software community development model is more powerful.<BR/><BR/>So, if we succeed to cooperate in a community the development is much more sustainable. And generally said a community can never go bankrupt :-) a company could (even though I do not think MySQL/Sun would in the foreseeable future).<BR/><BR/>Of course there are many things to discuss how communities can be more open, inviting and outreach more. Here it is sometimes easier if you have the more traditional structure of a company. I am curiously looking forward to see how things are continueing here :-)Unknownhttps://www.blogger.com/profile/09291585725493350108noreply@blogger.comtag:blogger.com,1999:blog-5552895785228669482.post-88354774151794889052008-12-01T13:19:00.000+02:002008-12-01T13:19:00.000+02:00MySQL Cluster _is_ carrier grade.MySQL Cluster _is_ carrier grade.Mark Callaghanhttps://www.blogger.com/profile/09590445221922043181noreply@blogger.comtag:blogger.com,1999:blog-5552895785228669482.post-82752214193559958592008-12-01T08:04:00.000+02:002008-12-01T08:04:00.000+02:00I do not understand why Marten Mickos hasn't been ...I do not understand why Marten Mickos hasn't been replaced yet by Sun management. He has little to no experience in running an open source software team, and continues to surround himself with marketing types, not execution specialists. <BR/><BR/>I have always known Sun to deliver carrier grade platforms, and it sure seems like Marten's in the way of having MySQL gain that status. Has anyone spoken to Sun management about this?Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-5552895785228669482.post-24747961340461701972008-12-01T07:21:00.000+02:002008-12-01T07:21:00.000+02:00Well written article for a sensitive topic Monty.W...Well written article for a sensitive topic Monty.<BR/><BR/>We know you guys fight daily at the source code front.<BR/><BR/>Keep up the good spirit !!Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-5552895785228669482.post-35074582833855898382008-11-30T19:39:00.000+02:002008-11-30T19:39:00.000+02:00About MySQL 5.1 being slower: As long as you don't...About MySQL 5.1 being slower: As long as you don't use logging to tables, MySQL 5.1 should be of about the same average speed as MySQL 5.0 (There are even some optimizations that makes some queries notable faster). I don't know of anything else that would cause a notable slowdown in MySQL 5.1. Please try to find out what could be the problem and post a bug report and the MySQL developers will try to find and fix the problem! In spite of everything we treat any bug that causes a crash or slowdown very seriously. The main problem just now is that the MySQL organization don't have a strong release criteria that defines when we can call something GA!Montyhttps://www.blogger.com/profile/06049512911785594864noreply@blogger.com