Slashdot stories can be listened to in audio form via an RSS feed, as read by our own robotic overlord.

 



Forgot your password?
typodupeerror
Cloud Databases Programming Python

Testing the MongoDB Global Write Lock Improvements 38

Posted by Unknown Lamer
from the heckling-haskell-hackers dept.
rick446 writes "I took some time to benchmark the global write lock improvements in MongoDB 2.0. From the article: 'MongoDB, as some of you may know, has a process-wide write lock... Per-database and per-collection locking is on the roadmap ..., but it's not here yet. What was announced in MongoDB version 2.0 was locking-with-yield. I was curious about the performance impact of the write lock and the improvement of lock-with-yield, so I decided to do a little benchmark, MongoDB 1.8 versus MongoDB 2.0.'"
This discussion has been archived. No new comments can be posted.

Testing the MongoDB Global Write Lock Improvements

Comments Filter:
  • Relevant (Score:5, Funny)

    by tibbetts (7769) <jason&tibbetts,net> on Monday January 02, 2012 @11:25PM (#38569294) Homepage Journal
    http://howfuckedismydatabase.com/nosql/ [howfuckedi...tabase.com] (Some NSFW language.)
  • by Anonymous Coward

    It's like this articles starts in the middle of a sentence and I can't tell what the hell is going on.

    OK, for starters, what the fuck is MongoDB? Just a single sentence or some mention would be helpful. Secondly, why is this front page material? It's just some crappy blog about some minor change to some product nobody uses, woopdeedoo.

  • Yea me too (Score:4, Funny)

    by Osgeld (1900440) on Monday January 02, 2012 @11:58PM (#38569390)

    I hate it when I am benching my mongo and it locks its yeild, quite painful =O

  • It's great that they improved the locking. I just hope they didn't compromise web scalability in the process.
  • Pushing locks down (Score:3, Insightful)

    by Anonymous Coward on Tuesday January 03, 2012 @12:33AM (#38569556)

    Congratulations. You matter enough to bother reinventing this wheel again. If you continue to matter for a meaningful amount of time you'll end up locking individual documents, or whatever you call them. Oracle called that 'row' locking. 15 years ago.

    • by FlyingGuy (989135)
      Mod + 1x10^1000
    • by steelfood (895457)

      I don't understand why they (and almost every other bit player in databases) are loading everything into main memory. I know disk is slow, but there are more efficient uses for main memory. Like creating a list of locked memory locations. The initial disk hit (which can be offset by intelligent caching and in this case, intelligent lookup) is probably not nearly as detrimental to the database's usability as the limitations from putting the entire database into main memory.

      As for locking, one of their partic

  • "MongoDB (from "humongous") is a scalable, high-performance, open source NoSQL database"

    LOL I think the word in parens they were looking for is "humours".

  • by Rogerborg (306625) on Tuesday January 03, 2012 @02:30PM (#38575486) Homepage
    For when you're too cheap to spring for a BerkeleyDB license, some amateur playing a decade of catchup gives you everything you'd ever need, as long as you don't need support, performance, stability or data integrity.
  • These comments seem up-to-date. That is if they were posted 5 years ago.
  • because i've saved so much time not having to write 500+ lines of DDL for this new web application in MongoDB compared to using a relational database.

If you're not careful, you're going to catch something.

Working...