> Does the error ordinarily happen often enough that you can be confident that
> the barrier option makes any difference at all? It could easily be something
> else, though the fact that you don't see the problem on OS X suggests a file
> system issue.
>
> The other thing you might experiment wi
On Oct 12, 2011, at 9:16 AM, Michael Ellis wrote:
> Hmmm, still not completely out of the woods. I've seen one instance
> of the error so far on one of the new systems. I ran "PRAGMA
> integrity_check" from a sqlite3 command line and saw a bunch of "page
> never used" reports. Then I ran VACUUM
Hmmm, still not completely out of the woods. I've seen one instance
of the error so far on one of the new systems. I ran "PRAGMA
integrity_check" from a sqlite3 command line and saw a bunch of "page
never used" reports. Then I ran VACUUM and re-ran the integrity check
which came back ok.
On Oct
> I'd turn on the ext4 barrier option just for the heck of it; see what happens.
>
> http://lwn.net/Articles/283161/
Thank you, Jonathan! That seems promising. I'm now testing it on two
systems. So far so good. Let's see if it holds up for a few hours.
I can't believe the Linux maintainers dec
4 matches
Mail list logo
