Reboot from GUI don't find USB to boot from

[peter_sm]

Hi,

 

When I stop the array and then do a reboot from the GUI, it will not boot again! the server are cycling trough to find any boot device ....... I then need to hit the powerdown button to powerdown the server, and then start the server, then it's find my boot device! very annoying.

 

Doing powerdown -r get into the same issue.

 

This have not been an issue before. I think I have this on b10 as well.

 

Anyone have any idea what causing this behaviour?  is not the PC getting a reset like a normal power down / power on when doing reboot?

 

My BIOS is up to date with the latest version since 8 months ago.

 

//Peter

 

 

syslog.zip

[saarg]

To me this sounds like a problem with your usb stick than a software problem.

I would try another usb stick to see if that works.

Do you have the usb stick in the boot menu when this happens?

[peter_sm]

I have tested 2 different USB, same issue. even on different ports.

 

So execute reboot it just cycle trough trying find boot device, but then power down and power on, it's found the boot device just fine.

 

One thing I saw today was after I have power down my server, I tested to send WOL , and to my surprise, the server started to boot ! :-) didn't know that was possible!

 

But the reboot issue is strange .......

 

//Peter

[peter_sm]

Do you have the usb stick in the boot menu when this happens?

In BIOS this USB is set to the boot device.

[sureguy]

This is an issue with your server.  Try doing the same thing with another bootable usb distribution

[peter_sm]

Hmm , my 2 different USB devices show same thing. I also moved these to different ports, no boot after execute "reboot" from unraid. But power off/on all good.

[garycase]

Have you made any hardware changes recently?    In particular, is the Br10i a new addition?

 

It sounds like a cold boot (from power off) is booting to the USB flash okay; but if it's a warm restart it's trying to boot from the Br10i, and doesn't find any bootable device.    If the BR10i initialization is taking a few seconds, this could explain why it doesn't do that when powering up ... but if it's already initialized, it may be attempting a boot on a restart.

 

If that's the case, a modification to the card's BIOS should resolve that.

 

 

[peter_sm]

No new Hardware is added to the server. And BR10 is an old card.

[garycase]

Very interesting.  With no new hardware added, this change in behavior is indeed very perplexing.

 

I've found that very strange behavior can sometimes be traced to out-of-spec PSU voltages; or to a failed BIOS battery.  Your motherboard's new enough that it shouldn't be the battery ... but they're only a couple bucks, so I'd try replacing it just to be sure.    If possible, the next thing I'd try is a different PSU.

 

 

[peter_sm]

Thanks garycase for input on this, but I don't have any other PSU.

 

Will look at the Br10 bios, and see if there is something to change? but doing a parity check right now.

 

//Peter

[archedraft]

hmmm I have noticed this happening on my server lately as well. I stop the sever from the gui and then reboot and it just cycles on the startup boot / bios screen but never finds the USB boot option... I also tried a different USB stick and reformatted my original USB stick but the problem persists.

 

I have to hold the power button to shut off the computer to stop the cycling and then it will find the USB stick and boot.

[peter_sm]

Same as me, Wonder if this is related to this B10? Can't believe is my server ;-)

[archedraft]

I know it's a recent problem with my server so it would seem like to much of a coincidence for both of us to all the sudden have the same issues and not be software related.

 

I believe that you are still using kvm  w?th passthrough correct?

 

If so the some common items our server share are:

 

1. KVM VM's

2. Video passthrough

3. Modified Syslinux.cfg

 

I'm not sure it would help but maybe getting the tail end of the syslog on a reboot would give us more to work with. Any thoughts?

[peter_sm]

We are in the same boat.

 

//Peter

[garycase]

A simple way to determine if this is somehow related to Beta 10 is to (a) save the entire contents of your flash drive; and (b) reload the flash drive with an earlier Beta that you know worked -- or even better with the v5 stable release;  and then see if the problem is gone.

 

That won't tell you WHAT is happening ... but it will certainly tell you that it's not a hardware issue.                                                                                               

 

[peter_sm]

I'm middle of changing / upgrade disk, can you do it? it would would be great! OR if LT can verify it easier with different releases! I have PM:ed Jonp about this behaviour.

 

//P

[archedraft]

A simple way to determine if this is somehow related to Beta 10 is to (a) save the entire contents of your flash drive; and (b) reload the flash drive with an earlier Beta that you know worked -- or even better with the v5 stable release;  and then see if the problem is gone.

 

That won't tell you WHAT is happening ... but it will certainly tell you that it's not a hardware issue.                                                                                             

 

I agree this would be a great way to test. My only worry is what might happen to my VM settings if I go to an older version of unRAID 6 which did not support VM's yet. Also some of my plugins for the VM's are specifically designed for the latest beta and who knows what might happen if I downgrade. What I might try is to disable my VM's from autostarting and turn off the plugins and see if that fixes anything. If I still have issues booting then I could try downgrading and the plugins will be disabled so the settings shouldn't be affected. I won't be home to test for a few hours so hopefully JonP might have some suggestions before then.

[garycase]

My only worry is what might happen to my VM settings if I go to an older version of unRAID 6 which did not support VM's yet.

 

That's why I suggested copying the entire flash drive first => you can experiment with an earlier version to see if the problem persists; and then simply restore the contents of the flash drive.    In fact, I'd just go to the stable v5.0.5 version, and only assign your data drives -- no cache or parity -- and simply see if the system then reboots okay.    Purpose of this test isn't to FIX the issue ... it's to TEST whether or not it's a hardware issue.    After you test it, simply restore the original contents of the flash drive and reboot to your current version.  Without setting a parity drive, and without any writes to the array during the v5 test, everything should be fine.  If v5 wrote anything at all, a correcting parity check after you're back to v6 will ensure parity is restored to all-okay.

 

... Normally I'd think this problem would be completely unrelated to the OS, since the issue is at boot time, and apparently the OS isn't even being started.    The only reason I have any doubts is that two of you are now having the same issue, and you're both using the same Beta version.

[archedraft]

That makes sense. I will test with version 5 when I get home and report back.

[archedraft]

UPDATE:

 

I put unRAID 5.0.5 on the flash drive and then assigned the data drives only. I then performed a stop array and reboot from the web GUI and it rebooted like to should to the blue unRAID screen. So I think it is safe to rule out hardware related issues. It is also safe to say that it has something to do with unRAID 6b10a.

 

I then booted into safe mode on 6b10a (which means no plugins or modified syslinux.cfg) and then rebooted the server and it went into endless reboot mode. It is looking more likely that this is a bug in unRAID 6b10a and not a problem with a plugin or the modified syslinux.cfg

 

Here is a screen shot of unRAID being shutdown:

 

I have attached a full syslog. JonP / Tom please advise on how to help future test this bug.

syslog-2014-10-13.txt

[jonp]

Ok guys, sorry for not replying sooner, but this is a weird one and not a lot of information to go on (syslog wasn't overly helpful as well).  Clearly this is an odd behavior, but unless I can find a way to recreate it, we're going to be spinning our wheels here.

 

Here's what I'd like to see from each of you having this issue:

 

1)  Please boot up unRAID 6 beta 10a into safe mode (non-Xen) and attempt a reboot, see if it work.

2)  Post a copy of your current syslinux.cfg so I can see what additional kernel parameters are being used.

3)  Please record a short video of the looping reboots so we can witness the actual experience.

4)  Please take pictures of your motherboard BIOS settings (as much as you can).

 

This is the best way to start diagnosing.  Start with #1 from above.

 

- Jon

[garycase]

Nice to eliminate the hardware ... but REALLY strange that this is OS-related in some way, since the boot process is "pre-OS" -- and the problem is the system isn't recognizing a boot device.

 

One interesting "coincidence" :  Both of you guys have  AsRock Pro4-M motherboards ... one with the Z77 chipset, one the H77.    But it's still perplexing that the reboot won't work; but a shutdown and then reboot works fine.    It's as if the OS is somehow setting a flag to change the boot order on a restart; but it gets cleared by powering off.  The reboot loop you're seeing with a "Safe Mode" boot of v6 makes more sense -- at least the system is finding the boot device and trying to restart ... clearly there's a bug that's causing this loop; but it's not in the BIOS boot selection area.

 

... hopefully JonP will see something in your lot that provides a clue as to what's happening.

 

[bungee91]

I've noticed this same condition lately and (for what it may be worth) also have an ASRock board, but not the same as it's a Z97 chipset.

I found it odd to say the least, and never had this issue prior to lately.

The weird part is that I don't get to a default "no boot device, press a key ..." message.

It seems like what is happening is that it attempts to boot from USB, fails, and then loops in this manner.

 

If this is the condition it would make a lot of sense on Unraid/USB boot device being the culprit and not the BIOS/post process.

 

A complete power off fixes this condition for me also.

 

 

[peter_sm]

Ok guys, sorry for not replying sooner, but this is a weird one and not a lot of information to go on (syslog wasn't overly helpful as well).  Clearly this is an odd behavior, but unless I can find a way to recreate it, we're going to be spinning our wheels here.

 

Here's what I'd like to see from each of you having this issue:

 

1)  Please boot up unRAID 6 beta 10a into safe mode (non-Xen) and attempt a reboot, see if it work.

2)  Post a copy of your current syslinux.cfg so I can see what additional kernel parameters are being used.

3)  Please record a short video of the looping reboots so we can witness the actual experience.

4)  Please take pictures of your motherboard BIOS settings (as much as you can).

 

This is the best way to start diagnosing.  Start with #1 from above.

 

- Jon

 

 

I have all disk formatted as XFS FS, what is the best older beta to try with? any links to b9 and B10?, I think the "reboot" was OK on one of these.

 

//Peter

 

[jonp]

Don't try an older beta yet.  Try the steps I've outlined and share the info.  thanks!!