Start a Conversation

Unsolved

J

9 Posts

1381

December 3rd, 2020 12:00

MD3620, unable to access via serial or IP

Background:

I've had this MD3620 in production for quite a while and it's worked very well. All the drives in it were about 5-6 years old and we were ready to upgrade our test environment so I decided to use this and replace all the drives with some 1 TB drives. At this point the PowerVault was working just fine. I had all the drives sitting in array but not inserted as I was waiting for something and had planned on doing that the next day.

Turns out that night we had a circuit breaker in the UPS that was powering that MD3620 go out while all the drives were not inserted. Booting it back up and I get "client_corrupt_db_detected_on_alt." Since this is all "new" and doesn't contain any data that isn't a big deal, i can just wipe and restart. Except I can't.

What's happening:

While keeping the original RAID controllers (0M6WPW) in it, I just get that error. I try to break using Putty so I can access the shell and it doesn't work. At first it still had my management IPs but MDSM just told me the config needed to be restored and wouldn't let me do anything with it. No matter what I do, it will not break to let me enter any commands.

I also tried the SMcli and the best I was able to get was this:

I ran: SMcli.exe 10.20.60.15 -c "show storageArray healthStatus;" and that gave me:

Syntax check complete.

Executing script...

The following failures have been found:
Storage Array in Recovery Mode
Storage array: Unnamed
Failure number: 434

Script execution complete.

SMcli completed successfully.

I then tried several different attempts at clearing the config but everything either said it was an invalid command or was unable to execute because it was in lockdown mode.

Coincidentally, we have another MD3600 that uses the same cards and it had a failed card. I bought 2 replacements so I could have a spare. At first I left the top card in the problem MD3620 and then it just wrote the config the the newly inserted one and put me back at square one.

I now have the second new controller ONLY inserted and was never added alongside one of the originals. This time it boots just fine BUT I can't do anything with it. If I set it up to get a DHCP address i get a message that it's not a permanent address but it receives one and I can ping it just fine. However, I am not able to access it via the MDSM. Via the console cable I try and do a break but same thing - it won't break where I can access the cli.

Question:

How can I get this thing back online?

Moderator

 • 

7.6K Posts

December 4th, 2020 15:00

Hello jeremy_lumos,

Are you trying to access the controller while it is trying to boot or are you waiting till after it has booted and is showing the lockdown message?

December 7th, 2020 14:00

I've got a serial cable connected to it and that's the only way I can currently connect to it. It is not fully booting as it's giving the error messages during the boot process

Moderator

 • 

7.6K Posts

December 7th, 2020 15:00

Hello jeremy_lumos,

When you do the control break are you hitting esc or hitting S?  you need to hit esc so that you can get access to the VxWorks login.

December 11th, 2020 12:00

Tried hitting esc but that didn't do anything. I get the same result

Moderator

 • 

7.6K Posts

December 11th, 2020 14:00

Hello jeremy_lumos,

If you are not able to access vxworks login for either controller, then you will need to replace the controllers. 

December 17th, 2020 10:00

These are new controllers. That's why i created the ticket. I've already replaced the controller and on the new controller I cannot get it to break to enter the terminal.

Moderator

 • 

7.6K Posts

December 17th, 2020 12:00

Hello jeremy_lumos,

What settings are you using to connect to the serial ports?  Are you able to see the controller trying to boot?  Are you trying this on both controllers & you are not able to get to VXshellusr on either controller?

January 4th, 2021 13:00

sorry for the delay, went on a much needed break. Yes - i can see it boot. It gets through most of the boot process then tells me that the database is corrupt. That's not a big deal since it's all new drives and i was trying to set it up again anyway. But it won't let me bypass that to actually overwrite the database or do anything else with it.

January 4th, 2021 14:00

I'm running this again now. I get past the POST just fine, i don't see any errors other than the drives being marked as incompatible but i think that's just because they are all new drives. I don't think that's part of the issue. I then get this message: 01/04/21-21:49:08 (tRAID): WARN: ********************************************************* 01/04/21-21:49:08 (tRAID): WARN: ********************************************************* 01/04/21-21:49:08 (tRAID): WARN: ** ** 01/04/21-21:49:08 (tRAID): WARN: ** WARNING!!!!!! ** 01/04/21-21:49:08 (tRAID): WARN: ** ** 01/04/21-21:49:08 (tRAID): WARN: ** This controller is not running the correct firmware ** 01/04/21-21:49:08 (tRAID): WARN: ** release. SOD is now suspended. In order to prevent ** 01/04/21-21:49:08 (tRAID): WARN: ** any configuration loss, you must download ** 01/04/21-21:49:08 (tRAID): WARN: ** x8.20.xx.xx release. ** 01/04/21-21:49:08 (tRAID): WARN: ** ** 01/04/21-21:49:08 (tRAID): WARN: ** ** 01/04/21-21:49:08 (tRAID): WARN: ********************************************************* 01/04/21-21:49:08 (tRAID): WARN: ********************************************************* I plug it into a switch and it does receive a DHCP address, however, when trying to access it via the MDSM it says "the specified device was not accessible. the storage array may not be connected properly or you may need to restart the host agent software. Refer to the online help for more information".

Moderator

 • 

7.6K Posts

January 4th, 2021 16:00

Hello jeremy_lumos,

How many drives are loaded in your system?  When you get to that point in post if you hit Crtl+ Esc do you get the prompt asking for username & password?

January 5th, 2021 12:00

Ok I've been trying to get it to break so i can get to that shell for about 2 months now and nothing i do works. I just discovered that Ctrl + Esc activates a Microsoft shortcut that cannot be disabled for the windows key. I then booted to a Linux distro and have putty running with the 115200 speed and even that won't do the break in the boot. There has to be another way to break this? I just spent over $2k on drives for this and now it's basically a brick that can boot but not do anything else.

Moderator

 • 

7.6K Posts

January 5th, 2021 17:00

Hello jeremy_lumos,

Here is what you want to do. When connected to the controller you want to

1.       Press Ctrl-Break

2.       Instead of hitting S press Esc

3.       You should get a vxWorks login

4.       Once logged in then you can run the clear lockdown command.

January 6th, 2021 07:00

This didn't work. I try ctrl + break and it still acts like nothing happened when I do that. It completely ignores any input from me

Moderator

 • 

7.6K Posts

January 6th, 2021 16:00

Hello jeremy_lumos,

You have all the orginal hardware that came with the system loaded in the system when you are trying this?  If that is the case then you should be able to access secure shell.  You are using the MD serial cable that came with your MD3620 or are you using a different serial cable? 

No Events found!

Top