• advexpwoo

    (@advexpwoo)


    Hello,

    Do you know if it is true that InnoDB tables can be more “sensitive” and crash.
    My large WordPress wp_posts and WP_users tables often crash and need to be restored.
    Could it be because they are InnoDB ?

    Server is reliable and I never had any problems for years before using WordPress. It has been checked by techs and everything is fine. Actually the same problem happens on 2 different servers having the same kind of WP installs.

    It happens only with with large tables, not on websites with few posts and users.

    Do you think that moving to MyIsam could help ?

    Also I found that on similar websites the same plugins sometimes have innodb tables and sometimes myIsam tables ? What could be the reason for a plugin to have different tables types on different websites ?

    Last, do you know how to find what can cause the crash of my tables ? I guess that it happens when something is being written ? But how to find exactly what it was doing when the table crashes ?

    Sorry if it sounds stupid, but I am trying to understand why the tables often crash few times every year (sometimes every 2 months).

    Thank you for your ideas.

    • This topic was modified 7 years ago by advexpwoo.
Viewing 2 replies - 1 through 2 (of 2 total)
  • Hello @advexpwoo,

    Do you know if it is true that InnoDB tables can be more “sensitive” and crash.

    That sounds like my mechanic talking! Being an ancient fellow he’ll complain and moan about the “problems” with any modern car with automatic transmission and all their fancy electronics which he can’t make a head or tail of.

    Even if true, note that being “sensitive” to crashes is not the same as “causing” the crashes. It just means MyISAM may continue to function even in the mist of errors, whereas InnoDB may give up much quicker.

    Either way, you still need to find the real cause of the crashes and fix it.

    Also, InnoDB has distinct performance advantages, which is why beginning version 5.5.5, the developers building the MySQL software found it wise to switch the default storage engine from MyISAM to InnoDB.

    It has been checked by techs and everything is fine.

    That gave me a chuckle!

    If everything is fine, then it should NOT be crashing. Don’t you agree?

    Could it be that the “techs” don’t have the chops to know what’s really going on, or more likely they can’t be bothered to dig deeper?

    The fact that it keeps crashing, by definition, means everything is NOT fine.

    It happens only with with large tables, not on websites with few posts and users.

    I don’t know your definition of “large”, but I manage a client’s site with over 100,000 posts (it’s a news portal) with InnoDB and no issues at all. The site is not idle either, I just checked Google Analytics, and we made 300k UNIQUES last month. And by the way, all this is on a highly optimized but tiny $20/month Linode VPS!

    Also I found that on similar websites the same plugins sometimes have innodb tables and sometimes myIsam tables ? What could be the reason for a plugin to have different tables types on different websites ?

    By default, WordPress uses the database’s default storage engine. In other words, WordPress doesn’t pick a type… it just takes what it gets from the server. (This default may be coming from a server-wide setting, user setting, or may have been set at the time the database was created).

    While plugins can declare storage engine when creating new tables, if the same plugin is using different storage engines on different sites, it probably means the different database (powering the different sites) have different default storage engines.

    Also, note that depending on the application, it CAN BE (but that doesn’t seem to be your case) an optimization strategy to have different tables in the same database using different storage engines.

    Sorry if it sounds stupid, but I am trying to understand why the tables often crash few times every year (sometimes every 2 months).

    Tables can crash for any number of reasons, whether they are InnoDB or MyISAM (or any other storage engine).

    Rather than guesses and speculations, get your “techs” to check the server logs and tell you the REAL REASON why the tables are crashing. That is, if they’re doing proper logging to begin with!

    If you can provide some log data, we’ll be able to give you a more informed advice.

    Cheers!

    Thread Starter advexpwoo

    (@advexpwoo)

    Thank you so much for the great reply.
    I asked Dreamhost many times to check and they replied that they didn’t find anything.
    Is there something that I could check myself ?
    The website has only 5000 pages (actually nothing compared to yours).
    I am not expert but if you have some links to read and try to understand I will do my best.
    I am a bit desperate, do you think that moving to another server could help, if Dreamhost has no solution ? In your idea is it a server problem ?
    Thanks again, I really want to find why the tables are crashing.
    Cheers.

Viewing 2 replies - 1 through 2 (of 2 total)
  • The topic ‘Convert InnoDB to MyIsam because of tables crash ? How to find the reason ?’ is closed to new replies.