Random hardware failures will NOT only impact a single file. The only thing that I know that will consistantly fail at the same spot is s/w.
On Mon, 03 Jan 2005 16:32:31 -0800, Mark Maggelet <[EMAIL PROTECTED]> wrote: > Thanks Mike, > It's hardware raid, but I don't know the manufacturer. I could try to > find out but if the problem is the hardware there isn't much I can do > about it anyway without turning it into a big project. > > I'm not really expecting to solve this, I'm just hoping for some advice > on what the problem is most likely to be (raid,kernel or mysql) or maybe > there is a variable that I can tweak. > > The thing that makes me think it's mysql is that it always happens to > the same table (out of 300 or so). The table it happens to has the most > reads by far but not many writes. > > any other tips are appreciated. > thanks, > - Mark > > mos wrote: > > > At 12:37 PM 1/3/2005, you wrote: > > > >> Hi, > >> I have a big table (900k rows, 200M MYD, 200M MYI) with a fulltext > >> index on it. The table gets corrupted every 1 week or so and I have to > >> repair it. I've tried upgrading to newer versions of mysql 3 times but > >> the problem persists. I think it may have something to do with the > >> raid 1 mirror but I can't really tell. > >> > >> Does anyone have any ideas? > >> Thanks, > >> - Mark > >> > > > > Mark, > > This might help > > http://www.google.com/search?hl=en&q=mysql+raid+corruption&btnG=Google+Search&meta= > > > > You can also check the groups > > http://groups-beta.google.com/groups?q=mysql+raid+corruption&start=10&hl=en&lr=& > > > > > > You may want to try and narrow down the results by also > > supplying your Raid hardware/software that you're using. It could also > > be processor/OS related. > > > > Mike > > > > -- > MySQL General Mailing List > For list archives: http://lists.mysql.com/mysql > To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED] > > -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]