SOLVED Reboot/Shutdown sometimes don't work (Nightly 2014-10-24)

[Sir.Robin]

I have this issue on workstation, but not on vsphere 5.5. :)


Sent from my mobile using Tapatalk

 

[sef]

8G RAM is the minimum suggested for FreeNAS. I wouldn't fully expect that to make a difference, but can you try it?

But I will create a VM with that right now with those specifications.

Edited to add: Several reboots with upgrades and new boot environments, and no boot failure. Fusion Pro 7.0.1 with a 2-core CPU, 4G of RAM, 8G SCSI hard drive, and a 20G SCSI hard drive.

 

[joeschmuck]

There is definitely something going on with VMWare and FreeNAS, I wish I could find the smoking gun. Running in a VM really helps do some quick testing for Betas.

The 4GB RAM and increasing it to 8GB will make no difference but I do understand your point, 4GB isn't minimum spec.

 

[kjp4756]

I see that joeschmuck's vm host uses an i7 950. My mac running fusion has an i5 750. They're both first gen i series. Maybe that's what's causing this?

 

[joeschmuck]

Not sure but I what version of fusion are you running? @sef is running Fusion Pro 7.0.1 without issue.

 

[kjp4756]

Not sure but I what version of fusion are you running? @sef is running Fusion Pro 7.0.1 without issue.

7.1.0 Pro is what I have.

 

[sef]

I upgraded to 7.1 last night, and did not have any problems with it, either. My MBP is an i7, for what that's worth.

 

[sef]

I installed those updates and upon reboot the console stopped at "Timecounter "HPET" frequency 14318180 Hz quality 950". I power off the VM and turn it back on and everything is fine.

I will note that that is where both USB is probed, and the root filesystem is mounted.

I have absolutely no explanation for what could be going on here; especially with a VM, power cycling shouldn't have an effect!

(Somewhat interestingly, I have run into this issue with my physical FreeBSD systems, where it was due to not having put zfs_load=YES in the loader.conf file. While that would be interesting if it were the reason here, that wouldn't change by doing a reset or power cycle, so I am left mystified.)

 

[kjp4756]

I tried removing the usb controller from my vm and tried installing some updates. I still get the hang. One thing I did notice is that I don't have to shut down the vm completely and power it back on. I can simply chose restart from the virtual machine menu at the top of the screen and the machine boots normally.

 

[sef]

For VMWare, restarting and power cycling the VM should be the same, I would think. I could be wrong.

I can't think of any reason that the kernel would be loaded, but then the system would hang at that point. I assume you don't see any messages about it waiting for the root filesystem?

 

[kjp4756]

No I'm not seeing any messages about waiting for root. My screen is exactly like the screenshot joeschmuck posted.

 

[joeschmuck]

Something I noticed was when I do a cold boot it will take longer from the start of the bootloader before text starts scrolling across the screen vice a warm boot where it appears maybe the system files could be cached and the text on the screen starts scrolling much sooner.

I will have to test out the restart on the VM to see if that makes a difference on my system.

EDIT: The VM Reboot does not work as I hoped, it only works after the system has crashed.

 

[kngpwr]

Disclaimer: This is my first forum post but I've been finding answers in here for years. I am a storage engineering professional during the day and a home FreeNAS user.

With that said, I'm encountering basically the same issue(s) with my new home setup. I know might make CyberJock grab his megaphone but here's my setup:

Dell FS12-TY C2100 Server
-Xeon L5520
-24GB ECC RAM
-LSI 2008 SAS HBA (For FreeNAS)
-LSI 1068E SAS HBA (For Vmware datastore)
-all the built in stuff
-lots of disks
-ESXi 5.5u2

My FreeNAS 9.3 nightly VM is setup with 3 cores, 12GB RAM, and DirectPath I/O (VT-d) to the LSI 2008 and the onboard Intel ICH10 SATA adapter. Since I've been testing the nightly builds I reboot at least daily. Each and every time the system hangs during boot with the last line displayed being:

Time counter "HPET" frequency 14318180 Hz quality 950

I can confirm the same behavior with VMware Fusion 7.x as well. In all case a reset results in a successful boot on the next try. This issue resembles an old bug with ESX and freebsd that was reported as fixed in ESX 5.0 patch 4. More info is here: http://kb.vmware.com/selfservice/mi...nguage=en_US&cmd=displayKC&externalId=2021185. Prior to that the work around was to enable the "ACPI Safe" timer instead of the HPET. I have not tried this work around yet since I found it just before making this post. After reading the thread I wanted to share my experiences and offer to help get this resolved if any of the devs are interested. I've added a comment to bug #6585 as well.

My experience makes me throw my first dart at vmware for being at fault but I have no evidence to support that. Its just where my spidey-sense is tingling today. How can I help?

EDIT: Just for fun I tested FreeNAS 9.1.2.8 and it reboots fine each time I try. Make of that what you will.

 

[sef]

As part of my test last night, I created a new VM on Fusion Pro 7.1, using the FreeBSD 64-bit template, with 4G of RAM, 2 processor cores, 1 8g SATA drive, and 1 20g SATA drive.

I've just done four reboots on it in a row -- two via the console menu, two via the web ui. No problems with any of them.

 

[kngpwr]

So I just upgraded from Fusion 7.0.1 to 7.1. Making a new vm similar to what you said reboots fine with hardware version 10 or 11. I've set the hardware to version 10 so it will work on ESXi 5.5u2 and I'm uploading it to the server. The previous machines were not "built" using Fusion 7.1. I'll update the thread when its done. Hopefully we are starting to zero in on something useful.

 

[kjp4756]

I'm beginning to see a pattern here.

1. I have i5 750 and experience the reboot problem
2. joeschmuck has an i7 950 and experiences the problem
3. kngpwr has a xeon l5520 and has the problem

All of these cpus are from around the same time.

Sef, what exact i7 do you have? I'm going to bet it's either a 2nd or 3rd gen i7.

 

[kngpwr]

So I just upgraded from Fusion 7.0.1 to 7.1. Making a new vm similar to what you said reboots fine with hardware version 10 or 11. I've set the hardware to version 10 so it will work on ESXi 5.5u2 and I'm uploading it to the server. The previous machines were not "built" using Fusion 7.1. I'll update the thread when its done. Hopefully we are starting to zero in on something useful.

Uploaded the VM to my ESXi server and the problem is now there. So the fun part is going to be sorting out what setting in ESXi makes it break when it does not in Fusion 7.1.

 

[kngpwr]

I'm beginning to see a pattern here.

1. I have i5 750 and experience the reboot problem
2. joeschmuck has an i7 950 and experiences the problem
3. kngpwr has a xeon l5520 and has the problem

All of these cpus are from around the same time.

Sef, what exact i7 do you have? I'm going to bet it's either a 2nd or 3rd gen i7.

Just to add my info here my Macbook Pro running Fusion is Haswell based. The thing is it had the same problem running Fusion 7.0.

 

[sef]

Mine is i7-4850HQ

 

[cyberjock]

I have an i7-920 as well as a Xeon E5520 (or something really close to it.. same generation as the other CPUs). In fact, my board is the X58-USB3 (the same as joeschmuck). Maybe I need to dust it off and plug it back in for one more test. :P