How long until the apologists for Linux's nonsense fsync realize you can just spam fsync() (maybe even just sync()?) to eat the error and prevent processes that care from *ever* seeing it?
-
-
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Let’s see if I’ve got this right: Anyone running a sync command on Linux can cause all others on the system to fail to detect filesystem corruption when it happens. And it’s Meltdown that people were worried about?
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Oh yeah Linux is nonsense. Yeah. Right. Really?
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
I remember having to tell some database developers long time ago that rows have to be updated in a table and its indexes atomically!
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
I’m not sure I understand...is this issue specific to fsync() on Linux everywhere?
-
I believe so, I think there may be a couple of outlier distros but mainly, yes.
End of conversation
New conversation -
-
-
wtf.... another bit of brilliant design!
-
y'all don't get - just like Nature, Linux isn't intelligent design, it is a product of a blind watchmaker.
- Show replies
New conversation -
-
-
JFS is *FAR* worse as fsync success means nothing as the kernel disk cache never gets flushed fully until the file system is unmounted. I’ve discovered this to my horror and pain after data loss due to a power outage on system with large memory. FS not flushed in months.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
I don't think the BSDs are any better. Cached writes and the buffer system make it hard to return errors. That's why databases are suppose to use raw devices, not file systems.
- Show replies
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.