I'm having an ongoing problem with my parity drive going offline that I have been unable to resolve and am still troubleshooting but while I am at it I have noticed that whenever it goes offline the cache drive also ends up in a bad state.  The unraid web console (along with unmenu) show it online and I can browse the directory structure (maybe due to cache_dirs?) but when I actually try to read or write a file I end up with an error.

 

It is obviously not actually online so why would unraid tell me it is fine?  Is this normal behavior or have I found a new bug?

And here is my syslog if it helps.  Oops, looks like it is too big to attach.  http://dl.dropbox.com/u/17526/syslog.txt

Wow, I'm a bit surprised that nobody seems to know whether their cache drive goes offline when their parity drive goes online.  Just knowing if this is normal or not would help me troubleshoot.  As it is I am thinking that it may not be a case of the cache going down because the parity went offline, I believe it may be that both the parity and cache are on the same sata controller that may be unreliable.

 

I moved the parity drive to another controller and four hours later the cache drive posted i/o errors. I have now moved the cache and parity both to a different sata controller and will monitor the situation to see if that resolves it.

It looks like the problem was the Marvell Sata3 chip running two of the ports on my motherboard, the ports that both my cache and parity drive were run on.  The odd part is that it doesn't just die, it works for a while, sometimes up to 12-24 hours, enough for my parity drive to fully rebuild, then eventually dies.  Since nobody knows whether it is normal for the cache drive to go offline when the parity drive goes offline I will test it this weekend by pulling the power to my parity drive and will update this post with the results once I have confirmed.  So far I have been running for over 24 hours with no problems since switching back to the sata2 ports.

It looks like the problem was the Marvell Sata3 chip running two of the ports on my motherboard, the ports that both my cache and parity drive were run on.  The odd part is that it doesn't just die, it works for a while, sometimes up to 12-24 hours, enough for my parity drive to fully rebuild, then eventually dies.  Since nobody knows whether it is normal for the cache drive to go offline when the parity drive goes offline I will test it this weekend by pulling the power to my parity drive and will update this post with the results once I have confirmed.  So far I have been running for over 24 hours with no problems since switching back to the sata2 ports.

 

It is not normal at all.  Check to see if those SATA3 ports are running in IDE/compatibility mode.  I have never had my cache drive and parity drive go down at the same time.

Good to know, I really appreciate that.  Initially I was looking for a connection, thinking one drive going down was taking the other down but now it's looking like the connection is that Marvell chip.  I just verified the ports are and have been setup in AHCI mode.