Jekyll / Hyde Server Making Me Crazy

[refthimos]

I purchased a 12 bay unRAID server to use solely as NAS for a buy Boxee Box. With one drive used for the parity drive and the other 11 with 2TB drives, that would give me 22TB of storage.

 

I started with just 4 data drives for a total of 8TB of storage. I've run out of space and need to add drives. But my server does not want to accept any new drives. I'm stuck with just the parity drive and the 4 data drives.

 

Why do I call this the Jekyll / Hyde server? Because it works as NAS storage - start it up, and the files are all there and accessible on my network (both by my Boxee Box as well as my computer). So in that sense, it works great.

 

On the other hand, the server is all screwed up when it comes to the interface. When I first got the server, I would get the tower login prompt. Now I don't get it unless I press ENTER after the startup sequence, which ends with the following:

 

Starting Samba: /usr/sbin/nmbd -D

                        /usr/sbin/nmbd -D

 

So it wouldn't be a big deal hitting ENTER to get to the login prompt, but even once I get there, I type with my keyboard and nothing shows up on the screen. Then once I hit ENTER, the characters show up, but no matter how many times I type "root" and hit ENTER (with no characters showing up on screen until after the fact), I can't get to the command prompt.

 

Again, I could care less accessing the command prompt since my needs for this server are so simple (just NAS for my Boxee), but when I try to add drives, the same Mr. Hyde behavior rears its ugly head... OK get ready: I first tried adding 3 new drives into bays 6-8 on the server. This caused the server to fail to boot up at all. Hmm OK... I then tried just adding a single drive unto bay 6. OK, now the server boots up, but nothing shows up in the web management page (it just shows my 4 data drives). If I put the single new drive into bay 7, it won't start up. Same thing when I try drive bays 8-12.

 

Any ideas?

[HAVOC]

Syslog would be helpful in assisting with your predicament.

[refthimos]

Syslog would be helpful in assisting with your predicament.

 

Understood, and believe me, I went through the forum and FAQ and understand that a syslog would be ideal. But since I can't get to the command prompt, I'm not sure how to accomplish this.

[dgaschk]

Reformat the flash and reinstall unRAID. The data drives will not be affected. Shares will appear with default configurations.

[refthimos]

Reformat the flash and reinstall unRAID. The data drives will not be affected. Shares will appear with default configurations.

 

Well, tried reinstalling unRAID as you suggested.  Good news is, I did this successfully, and when I went to the "Devices" tab, I saw that the server recognized my 5 drives (1 parity and 4 data) and was able to assign them as I had when I first configured the server. Bad news is, even when the server booted up (with just the same 5 drives that worked before), it still hung up and the end of the boot process and failed to give me a command prompt. Undeterred, I decided to go ahead and try to add drives - but same thing as before - the server hangs up and fails to boot. It doesn't get to the blue boot screen - in fact it seems to fail at the "Initializing RAID Component" stage. Since I can't access the command prompt in order to generate a syslog, I took a picture of the screen where the boot process fails:

 

 

All 5 of the drives that work as well as the 3 that I am trying to add are WD20EARS drives - if I've read the materials correctly, I should not be messing with these drives despite all the posts re: them. This is what I did with the first 5 drives - I installed them in their bays and unRAID recognized them with no problem (as it did this evening). But thought it might be worth mentioning.

 

Any other ideas?

[ohlwiler]

I think you have a hardware problem. I have been there, sometimes they can be difficult to diagnose.

 

I would pull your four data disks and parity drive and put them aside until you can get your server to be reliable. Bring up the system with your new disks and exercise it. Your problem could be memory, SATA controller card, power supply, hard drive, motherboard, flash drive or other. If you can recreate your problem, then start swapping parts. I created my first test server with parts I purchased trying to diagnose a stability problem.

 

Unraid can be incredibly reliable and incredibly frustrating when it is not. Good luck.

[Rajahal]

So it wouldn't be a big deal hitting ENTER to get to the login prompt, but even once I get there, I type with my keyboard and nothing shows up on the screen. Then once I hit ENTER, the characters show up, but no matter how many times I type "root" and hit ENTER (with no characters showing up on screen until after the fact), I can't get to the command prompt.

 

Are you using a wireless keyboard?  They usually work, but I've seen some that are spotty.  If you have a wired keyboard (USB or PS/2), that would be best.

 

I first tried adding 3 new drives into bays 6-8 on the server. This caused the server to fail to boot up at all.

 

This is most likely caused by the new drives stealing boot priority.  This is a 'feature' of nearly every motherboard on the market - whenever you add a new hard drive, the motherboard will try to boot from the new drive.  The solution is simple - boot into BIOS (hold DEL as the server boots), go to the 'Boot' tab, choose 'Hard Drives', and make sure the USB Flash Drive is the first boot device.  You can then disable all other boot devices.  Finally, press F10 to save and exit.  The server should reboot into unRAID.

 

Hmm OK... I then tried just adding a single drive unto bay 6. OK, now the server boots up, but nothing shows up in the web management page (it just shows my 4 data drives). If I put the single new drive into bay 7, it won't start up. Same thing when I try drive bays 8-12.

 

Again, this might be the new drives stealing boot priority.  However, it could also be a hardware failure, or just a loose cable.  We'll need more info before we can make a full diagnosis.  Try the steps above and let us know how it goes.

 

Best of luck!

[dgaschk]

Are you sure the SATA card is set to JBOD?

 

How many drives currently?

 

Which PSU?

[refthimos]

Raj,

 

Thanks for the tips. Actually it's funny, I think I may have been one of your first customers at Greenleaf, if not the first (I dealt with Stephen). Anyways, I am using a wired keyboard using the PS/2 connection. And I checked on the boot priority issue - went into the BIOS and it still is set to boot from the USB drive.

 

dgaschk,

 

The server has 12 bays, with the top 5 being used (1 parity, 4 data). As for the other 2 questions, Raj might know better than me! :-)

[refthimos]

So to recap, drive 1 (parity) and drives 2-5 (data) work fine when in my server's 1-5 bays. Inserting drives 6-8 (additional data drives) into bays 6-8 (I tried every combination, eg all three drives in together, one drive only in, etc) causes the server to fail to boot up (it stalls out as captured in the picture above).

 

I tried reinstalling unRAID and the problem is the same.

 

I pulled drive 5 (confirmed working from existing configuration) from bay 5 (confirmed working from existing configuration) and the server started up fine, noting that drive 5 was missing.

 

I tried booting up the server with drive 5 in bay 6, and the server booted up fine, though of course I could not add drive 5 to the array because there was a drive missing from bay 6.

 

Since bay 5 works just fine with drive 5, I then tried adding drives 6-8 (one at a time) into bay 5 - each time the presence of one of these drives (6- rather than drive 5 in bay 5 caused the server to fail to boot.

 

Eventually I will need 4 more drives (in addition to drives 6- in order to take advantage of the available bays in my server, so seeing as somehow the new drives 6-8 don't seem to be playing nice with my server the way drives 1-5 do, I went ahead and ordered a Seagate Barracuda 2TB drive to see if this helps things, though I am not too optimistic about this, since drives 1-5 are the exact same model as drives 6-8 (WD20EARS).

[Rajahal]

dgaschk - The SATA card is a Supermicro AOC-SASLP-MV8 with firmware 3.1.0.21 (as seen in the screenshot) and INT13 disabled.  It is in it's default JBOD mode.

 

refthimos - Looking through our previous correspondence, your server suffered from water damage some time back.  Possibly the SATA card was damaged?  Given the odd behavior you are seeing, I think that is a likely culprit.  It is also possible that your power supply was damaged, and that it can no longer power all of your drives.  That seems less likely to me, but it could happen.

 

I'm going to send you an email with a few more support ideas.

 

- Raj / Stephen