In testing unRAID v5.0b9 I noticed two potential issues.  I have a parity drive and two data drives.  I filled the data drives with content, then I pulled one of them out.

 

1) unRAID reports that both are still present, even though one is not in the case.

 

2) The default option for parity checking is to "Correct any Parity-Check errors by writing the Parity disk with corrected parity." which if I understand correctly will re-create parity from the single remaining data drive, effectively "losing" all the data on the missing drive.

 

For the first item, a user may not always have the unRAID web interface up, so I suggest that the user be informed immediately (via email, if possible) that the drive is disconnected, so they can swap a new drive into the case.  For the second item, I suggest that the checkbox for correcting errors be un-checked by default.

 

 

1) unRAID should not be reporting that he missing drive is still present. It should be showing it as missing.

 

2) You did something wrong if you have the ability to do a parity check while a data drive is missing.

 

I do agree with the arguement about email notification. There is one or more add-ons that will do that. Also, unRAID will have a proper plug-in architecture one day in the near future (possibly 3 to 6 months) and I'm quite sure an email notification plug-in will arrive shortly after the architecture is in place. Possibly even other forms of notification, such as twitter or growl.

 

Peter

 

Can you elaborate on what I might have done wrong?  I'd like to start using the server, but I want to work out the kinks first.

 

(I'd post my syslog, however it's larger than the size of my Flash drive, so I get an error generating it.)

The unRAID webGUI does not auto refresh, so pulling a drive and looking at the main page without having refreshed it would not have told you anything.

 

unRAID itself will realize the drive is missing and simulate it's contents.

I guess what I'm saying is that I think earlier warning would be better.  As it stands, that drive could be down for weeks or months and I wouldn't know unless I fired up a browser and looked to see if any drives had errors.

Look up the email notification then. The add-on unMENU package gives a fairly easy way to enable email notification.

 

Peter

So are you saying the unMENU email notification would trigger when the drive fails?

So are you saying the unMENU email notification would trigger when the drive fails?

 

There are a lot of things in unMenu and its package manager.  Take a look through it and install the email notification package along with the others that are related to drive over-temp and the like.

Pretty soon there will be a new plugin for 5b that will give notifications for array/disk status, as well as SMART/Temp failures

 

In unMENU "Array Management" I'm told the parity drive isn't working (DISK_DSBL), and that the two data drives are OK, even though I know the parity is fine because I've simply pulled one of the data drives out.

 

In unMENU's syslog, I can see errors reported for both the parity drive and the removed data drive.

 

myMain shows a red X next to the parity and two green checks next to the data drives.

 

I'm new to unRAID, so maybe the answer is eluding me, but how am I supposed to know which drive failed so I can replace it?  Rebooting seems to be the only way to find out that the drive has failed, but I had intended to leave the server running 24/7.

I formatted the removed data drive and put it back in (same slot).  Now the array status is "Stopped. New parity disk installed." which isn't exactly what I was expecting to see.  The parity drive hasn't changed, just one of the data drives.

I formatted the removed data drive and put it back in (same slot).  Now the array status is "Stopped. New parity disk installed." which isn't exactly what I was expecting to see.  The parity drive hasn't changed, just one of the data drives.

what is FAR more likely is you do not know which of your drives is the parity drive, and which is the data drive. 

I've never seen unRAID get them mixed up, since it tracks them by model/serial number.

post a syslog, zip it up to make it small enough to attach.  It not small enough, use pastbin or similar and post a link.

 

Joe L.

Likely or not, it's the truth.  My parity drive is the 3Tb EZRX.  I've attached my syslog for your reference.  On lines 889-890 it says the parity drive (disk0) was replaced, however this was not the case.  The drive I removed was disk1.  I've verified this physically, just in case you still believe it's FAR more likely that I'm an idiot.

syslog-2011-07-24.zip

I formatted the removed data drive and put it back in (same slot).  Now the array status is "Stopped. New parity disk installed." which isn't exactly what I was expecting to see.  The parity drive hasn't changed, just one of the data drives.

Did you by chance start the array when the disk was unplugged?  If so, it forgot the model/serial number of the parity drive.  Re-installing it now makes it think it is a replacement.

 

(and no, you are not an idiot.... I might be, but I've seem more new users of unRAID lose track of their drives then you.  In your case, it is impossible to get it wrong, as the 3TB must be the parity disk.)

That has a ring of truth to it--I must have started the array with that data drive unplugged, and that confused the system.  I was also under the impression my mobo was configured for AHCI, but apparently it lacks that feature, which means my attempt at hot-swapping wasn't helping either.  I'm going to reset the server and start testing again from scratch.