2011/6/18 Lukas Czerner <lczerner@xxxxxxxxxx>: > Hi, > > so you're saying that you encounter I/O error on access(2) only with > Ext3/4 with journal. So given that you're checking the error count in > ext4_handle_error() which is called when I/O error happens I fail to see > how this helps your case. Am I missing something ? > Only when access(2) return "Read-only file system", hadoop will mark the disk as offline. For Ext4 no-journal mode, there is no jdb2 to set the file system as read-only when I/O error happens, so we set an threshold, when io error number reach this number, we change the filesystem to read-only. I use ext4_abort(), maybe it is wrong? > Also I do not understand how this is helpful at all ? Usually when we > hit I/O error we want to have predictable behavior set by the error= > mount option, but with this patch we have absolutely unpredictable > behaviour on errors, which is bad! Also we can end up with read-only > file system even when errors=continue has been set. > In ext4 without journal, when the disk drops, the fs can't be readonly. But in ext3/4 with journal, jbd2 will abort the filesystem, change fs to be read-only. So we don't care what kind of error happen, we just want to change fs to be read-only when there are too many errors > You can use atomic_t and get rid of the spinlock maybe ? > Yes, thanks > The name for this function should rather be inc_sb_error_count(). Thanks > I am not sure, but given that it it a "threshold" should not we trigger > it when we hit the threshold and not threshold+1 ? Thanks, I should use ">=" > Could you use better error message ? This does not say nothing about why > it happened. Something about IO errors count reached the threshold ? Yes, IO errors count reached the threshold, we need change fs to be readonly > Maybe you can use atomic operations and get rid of the spin_lock. spin_lock is just a "lazy approach" -- Wang Shaoyan -- To unsubscribe from this list: send the line "unsubscribe linux-ext4" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
