Okay, was going to swap a drive today and upon bringing up the management interface to shut things down foond that my parity disk was offline and showing 8 write errors. Now, I have no reason to believe that the disk is bad and suspect possible cable issues. However no way will I swap in another disk with Parity not right (duh). So, I took the array offline to clear the statistics (no longer manually clearable through interface it seems). Well, upon stopping the system my Parity drive is now seen as "new". I started the array and it's currently writing to the Parity drive as if nothing happened Is that correct? It didn't recognize that this was a drive it had marked "bad" like I think it used to.

 

Also odd - upon starting the array the two drives I lost previously, Disk8 and Disk11 showed 12 writes during mounting instead of the 1 write that all of the others saw. Any ideas why that might be? Is that normal and expected and okay or is something goofy going on? Below is what I'm seeing right now...

 

Perhaps I'm just paranoid but umm, yeah I'm a little gunshy right now!

 

Sounds like it got 'disabled'.  Yes please check cabling.  You can see extra writes during a mount if some transactions get 'replayed' due to unexpected shutdown of the sytem - did that happen?  i.e., non-clean shutdown: reset, power off, etc.

 

The button to clear the Statistics was moved to the Settings page.

The system was never shutdown so no the cached writes shouldn't be an issue. Cabling seems okay but those upper two drives have given me problems before. If I have to power it down I think I may swap oout that cable just to be sure.

 

Something else I noticed - as you can see in the screen shot parity speed was low. That was RIGHT after it started so I assumed it would speed up as it usually does. Well, some time later it's still stuck in the 1.4meg range instead of the 12-13 that it usually is. I know that you had changed some things in Beta4 to solve this and it had seemed to work on mine but for some reason mine has slowed down again. Syslog shows nothing untoward so I'm confused. Any suggestions for troubleshooting this? Hopefully this is a completely different issue than previously as that seemed to have been solved.

I would suggest that you completely review your cabling.  If you have to, buy all new high quality cables.  DO NOT use the rounded cables, get the flat 80-conductor ribbon cables.  High quality flat ribbon cables are not expensive.

 

You're correct about parity speed not being accurate iniitially - this is because of all the i/o activity associated with mounting the drives.

 

If still slow "some time later", this is an indication of i/o problems and/or bad hard drive.  Nothing was done in -beta4 to "improve" this - changes in beta4 to were to attempt to provide better recovery from dma/irq timeout problems - which is probably not what you're seeing.

Okay, response is also VERY slow so something is certainly going on. I have round cables only on the bottom most cables - flat cables simply don't reach unfortunatly. I will halt the Parity, shut it down, and check things over to see if anything weird is going on - I'll swap out the suspect cable up top too. Strange that it has been fine but then gotten weird so we'll see what it does when I bring it back up.

Cable swapped and system rebooted. Am seeing 14.9meg right now. Not sure if it was the reboot or the cable but it's rolling. Nothing came up goofy either - always a good thing!

its the round cable... i had a bunch when I was going through my whole 20 page trouble.... i noticed the moment I swapped out my round to ribbon, things were moving much better. Tom, you should add something to the web page stating round cables suck, save your money and buy the ribbons 

A warning to not use round cables should be added to the docs. 

Umm, I didn't swap out any round cables. I am aware that they violate the standard blah blah but there were no flat cables of a length that would reach unless I wanted my drives mounted with their asses hanging out on the bottom slots. The cable I've swapped out was an odd yellow color, not sure where it came from. Standard length and was in the top most slot attached to the motherboard. In the past I've occasionally had that top drive or two not come up after a reboot. When the Parity drive came up with some errors the cable was suspect and I've been meaning to swap it anyway so... However with my previous issues the way it came up after a reboot was weird as I've been used to the drive being persona' non grata' after it's had errors. Actually prefer it this way Usually it's a cable issue and not a bad drive so this worked out fine... Swap in the new bigger drive tomorrow night perhaps - today's just been too long a day to mess with it!

don't forget to cut up that bad cable and toss it !!

 

 

 

No time again today to upgrade <sigh> Yeah, usually I whip out a Sharpie and write BAD on anything hosed like cables, memory, CPUs, or HDs. Some things come in handy to keep around even when bad.

Yikes, I've been running round cables for 6 months (no problems yet), evern 24 inch (maybe a 36).

 

OK, advise time guys:

A. Replace them now

B. If it ain't broke, don't fix it

B.

B but try to keep lengths as short as possible. IMO there's NO need for a 36inch cable in the CM case.