This article discusses an issue of a computer getting stuck at boot with the message “Scanning and repairing drive 9% complete” with chkdsk hanging at 9%.
The hypervisor was 12 months over-due for a BIOS update. Updating the UEFI should be simple enough, however SuperMicro have a nasty habit of clearing the CMOS during BIOS updates. Why most other OEM’s are able to transfer settings and SuperMicro insists on not is one of only a few gripes that I have ever had with the firm. Yet it is a persistent one that I’ve had with them going back to 1998.
The Fault
After the successful update, I reset the BIOS to the previous values as best I could recall. Unfortunately I also enabled the firmware watchdog timer.
SuperMicro’s firmware level watchdog timer does not operate as you might expect. It requires a daemon or service to be present within the running operating system that polls the watchdog interrupt periodically. If the interrupt isn’t polled, the firmware forces a soft reboot. Supermicro do not provide a driver to do this for Windows, although their IPMI implementation can do so.
After 5 minutes from the POST the hypervisor performed an ungraceful, uninitaited reset. Following the first occurrence I assumed it was completing Windows Update. Subsequent to the second, I was looking for a problem and after the third (and a carefully placed stopwatch) I had a suspicion that I must have turned on the UEFI watchdog.
I was correct and, after disabling it, the issue was resolved.
This particular hypervisor has SSD block storage for VMs internally and large block storage for backup via an external USB 3.1 enclosure – a lot of it. Without giving it any thought, I told the system to
chkdsk <mountPoint> /F
Note that this does notinclude the /R switch to perform a 5 step surface scan. I told chkdsk not to dismount the volume, but to bundle all of the scans together during the required reboot to scan the C:. Doing it this way meant that I could walk away from the system. In theory this would mean that when chkdsk finished, it would rejoin the Hyper-V cluster on its own and become available to receive workloads.
… and restarted.
Scanning and repairing drive 9% complete
chkdsk skipped the SSD storage as it is all configured as ReFS. Under ReFS, disk checking is not required as it performs journaling activities in the background to preserve data integrity. Unfortunately, the external backup enclosure volume was NTFS. It would be scanned – and it was also quite full.
The system rebooted, and sitting at the intermedia chkdsk stage of the NT boot process. It zipped through the SSD NTFS boot volume in a few seconds, before hitting the external enclosure. Within around 5 minutes it had arrived at the magic “9% complete” threshold.
1 hour, 2 hours, 4 hours… 8 hours. That turned into 24 hours later and the message was still the same.
Scanning and repairing drive (F:): 9% complete.
Crashing the chkdsk
The insanity of waiting over 24 hours had to come to an end and I used IPMI to forcefully shutdown the server.
After a minute or two, we powered back on. To be met with a black screen of death from Windows after the POST.
The c:\pagefile.sys was corrupt and unreadable. Perform a system recovery or press enter to load the boot menu. On pressing enter, the single option to boot Windows Server 2019 was present, and, after a few moments. Windows self-deleted the corrupt pagefile.sys, recreated it and booted -to much relief.
I then ran
chkdsk c: /f
and rebooted, which completed within a few seconds and marked the volume as clean, with no reported anomalies.
The Windows System Event Log contained no errors (in fact as you might expect, no data) for the 24 hour period that the server had been ‘down’. The were no ‘after the event’ errors added to the System log or any of the Hardware or Disk logs either. for all intents and purposes, the system reported as fine.
Trying chkdsk for a second time
I decided to brave running chkdsk on the external enclosure again. Initially in read-only mode
chkdsk F:
Note the absence of the /F switch here.
It zipped through the process in a few seconds stating
Windows has scanned the file system and found no problems. No further action is required.
Next I ran a full 3-phase scan
chkdsk F: /F
Again, it passed the scan in a few seconds without reporting any errors. So much for the last 24 hours!
Analysis
The corruption in the page file indicates that Windows was doing something. The disk array was certainly very active, with disk activity visible (via LED), acoustically and via data from the power monitor on the server all confirming that “something” was happening. Forcibly shutting down the system killed the page file during a write. Had been a 5-step chkdsk F: /f /r scan I could understand the length of time that it was taking.
With chkdsk /f /r – assuming a 512 byte hard drive – the system has to test 1,953,125,000 sectors for each terabyte of disk space. Depending on the drive speed, CPU speed and RAM involved it isn’t uncommon to hear of systems taking 5 hours per-terabyte to scan. This scan was not a 5-step scan, just a 3-step. A live Windows environment could scan the disk correctly in a few seconds.
Resources were not an issue in this system. Being a hypervisor, it had 128GB of RAM and was running with 2018 manufactured processors.
My suspicion is that the problem exists because of a bad interaction between the boot level USB driver and the USB enclosure. The assumption is that Windows fell into either a race condition or a deadlocked loop. During this fault, chkdsk was genuinely scanning the disk and diagnostic data was being tested in virtual memory (i.e. in the page file) but it was never able to successfully exit.
The lesson that I will take away from this experience is that unless it to avoid using a boot cycle chkdsk to perform a scan on a USB disk enclosure.
When you are designing your storage subsystem. On modern hardware, you will often be asked to choose between formatting using 512 byte or 4K (4096 byte) sectors. This article discusses whether there is any statistically observable performance difference between the two in a 512 vs. 4K performance test.
NB: Do not get confused between the EXT4 INODE size and the LUN sector size. The INODE size places a mathematical cap on the number of files that a file system can store, and by consequence how large the volume can be. The sector size relates to how the file system interacts with the physical underlying hardware.
Sector Size selection on QNAP QTS 4.3.6
Method
A QNAP TS-1277XU-RP with 8x WD Red Pro 7200 RPM WD6003FFBX-68MU3N0 drives running firmware 83.00A83 were installed with 8 drives in bays 5 – 12
Storage shelf firmware was updated to QTS version 4.3.6.0923, providing the latest platform enhancements
A Storage Pool comprising all 8 disks in RAID 6 was configured, ensuring redundancy
A 4GB volume was added allowing QNAP app installation so that the systme could finish installing
The disk shelf was rebooted after it had completed its own setup tasks
RAID sync was allowed to fully complete over the next 12 hours
Two identical 4096 GB iSCSI targets were created with identical configurations apart from one having 512 byte and the other 4k sector sizes
SSD caching was disabled on the storage shelf
2x10Gbps Ethernet, dedicated iSCSI connections were made available through two Dell PowerConnect SAN switches. Each NIC on its own VLAN. 9k jumbo frames were enabled accross the fabric
A Windows Server 2016 hypervisor was connected to the iSCSI target and mounted the storage volume. iSCSI MPIO was enabled in Round Robin mode. Representing a typical hypervisor configuration
The two storage LUNs were formatted with 64K NTFS partitions (recommended for dedicated VHDX volumes)
A Windows 10 VM was migrated onto each of the targets and the test performed using Anvil’s Storage Utilities 1.1.0.20140101. The VM had no live network connections. The Super Fetch and Windows Update services were disabled, preventing undesirable disk I/O. The VM was not rebooted between tests, had no other running tasks and had been idling for 6 hours prior to the test
No other tasks, load or data were present on the storage array
512 vs. 4K Performace Results
The results of the two tests are shown below.
Anvil Storage Utilities Screenshot with 512bytes results
IOPS
512 byte
4K
4K Diff
4K Diff % +/-
Read
Seq 4MB
417.45
403.3
-14.15
-3.51
4K
3001.56
3164.56
163
5.15
4K QD4
6021.45
6006.7
-14.75
-0.25
4K QD16
24228.16
24062.61
-165.55
-0.69
32K
2742.39
2807.47
65.08
2.32
128K
2628.86
2620.8
-8.06
-0.31
Write
Seq 4MB
233.2
230.79
-2.41
-1.04
4K
2090.79
2165.45
74.66
3.45
4K QD4
5976.18
5983.65
7.47
0.12
4K QD16
8254.84
7874.67
-380.17
-4.83
Analysis and Recommendations
The results show that there is little difference between the two. Repeating the tests multiple times showed that the figures for both the 512 byte and 4K LUNs are within the margin of error of each other. A bias towards 512 byte was consistently present, but was not statistically significant.
The drives in the test disk array are 512e drives. 512e is an industry transition technology between pure 512 byte and pure 4K drives. 512e drives use physical 4K sectors on the platter, but that the firmware uses 512 byte logic. A firmware emulation layer converts between the two. This creates a performance penalty during write operations due to the computation and delay of the re-mapping operation. Neither sector size will prevent this from occurring.
My recommendations are
If all of your drives are legacy 512 byte drives, only use 512
Should you intend to mount the LUN with an operating system that does not support 4K sectors. Only use 512
In situations where you have 512e drives, you can use either. Unless you intend to clone the LUN onto 4K drives in the future, stick with 512 for maximum compatibility
Never create an array that mixes 512 and 4K disks. Ensure that you create storage pools and volumes accordingly
WorkFolders allows you to perform policy based file HTTPS synchronisation between corporate servers and BYOD devices or teleworker devices. This article discusses a workaround to a problem where an anonymous sync error appears on a directory despite all of its contents synchronising successfully.
Outline of the Problem
Assume the following directory structure
C:\Users\CompanyUser\WorkFolders\Documents\WorkFolders Test
The following files/folders are present within WorkFolders Test:
WorkFolders Test\Problem Folder
WorkFolders Test\Problem Folder\File in Problem Folder.docx
WorkFolders Test\This file is OK.txt
Windows Explorer will display a Green circle with a tick for file/folder object that is synchronised and a Red circle with a cross for a faulted file/folder. After synchronising, the sync results will display as follows:
WorkFolders Test\Problem Folder [CROSS]
WorkFolders Test\Problem Folder\File in Problem Folder.docx [TICK]
WorkFolders Test\This file is OK.txt [TICK]
No errors are displayed in the Control Panel WorkFolders applet. There are no related errors in the clients WorkFolders Management/Operational Event Viewer logs. No relevant errors are present in the file servers SyncShare Operational/Reporting logs.
The parent folder shows that its sub-folder has a sync errorThe contents of the error’d folder are however correctly synchronised.
Analysis
Although WorkFolders is indicating that the issue is being caused by the “Problem Folder” directory. The issue is being caused by the “File in Problem Folder.docx”.
The following symptoms will be true:
Renaming “Problem Folder” will not fix the issue
Altering the filename of “File in Problem Folder” will not fix the issue
Changing the “File in Problem Folder.docx” file extension (e.g. to .txt) will not fix the issue
Opening and saving the “File in Problem Folder.docx” will not solve the issue
Moving “File in Problem Folder.docx” out of “Problem Folder” will clear the sync error, but the error will immediately migrate to the new location
Rebooting the client computer will not help
Restarting the server will not help
The file does not have any connected temp files or lock files associated with it in the client file system
There is nothing wrong with the file itself. It is not corrupt, pay-loaded with a virus or violating any policy. It is my (unproven) belief that the record for the file in the WorkFolders synchronisation database is corrupt. Performing any of the above steps will not alter the record in the WorkFolders client database, thus the problem cannot be ameliorated.
Fixing the problem
One you have identified the problem file(s). You can use one of the methods below to correct the error.
Save the file as a completely new file
Open the file in its associated editor (e.g. Microsoft Word for docx files)
File > Save as…
Save the file in its original location, but with a different file name (do not overwrite the original)
Delete the original file
Allow WorkFolders to re-sync
Rename the new file as required
This approach is easy for an end-user to perform, but can be very time consuming if you are troubleshooting a large number of such issues. It requires you to know which file is causing the problem in the first place.
Compression
Compress the file using Windows Compressed folders (Right click > Send to… > Compressed (zipped) file)
Delete the original file
Wait for the folder to re-sync and clear the error
Extract the original file from the zip back into the desired location
This method will create a new record in the WorkFolders synchronisation database and the error will not reappear. You can use the technique to fix an entire folder structure without having to first identify the problem file. It is also easy for an end-user to perform.
Move the file
Move the file outside of the WorkFolders monitored file system. For example, move the file to C:\ or into the Recycle Bin
Allow the original folder to re-sync and clear the error
Return the original file to its original location and allow it to re-sync
Again, this method will create a new synchronisation record. In a managed environment this may be harder for an end-user to perform due to permissions. It is however easier for an administrator to perform as you can cut/paste the entire file structure out and then back into the WorkFolders sync root.
If you use this method, remember to move it to a location within the same drive letter. If you do, the move will preserve permissions, file dates and will not physically copy the underlying data to the new location (just update the MFT).
In 2019, streaming your TV entertainment has become so popular that it is almost the norm. Systems such as Plex and Kodi create easy to understand, consistent and familiar cross-platform environments in which the whole family can consume media.
IPTV is an extension of such systems adding live broadcast playback and Personal Video Recorder (PVR) functionality. PVR adds the ability to watch, pause and record live TV; be it from aerial, satellite, cable or online sources. Many of these setups will use a local TV tuners plugged into a stand-alone media centre device. But what if you want to provide TV to multiple media centre appliances simultaneously? And what if you want that system to be a virtual TV streaming server instead of a dedicated streaming PC?
This article discusses how to create such a setup.
Why Virtualise?
I have spoken to a number of hardware and software providers in the course of this experiment. One thing that has been consistent has been their response: first laughter, followed by a dismissive “why would you want to do that?”.
Virtualisation is the process of taking what would be considered to be the work of a physical computer. Lifting it up and placing it – along with other workloads – onto another computer. This usually means that a single physical computer (a server) runs multiple, often different operating systems at the same time.
If you want to provide an always-on TV experience to multiple devices, then by definition this requires the TV server to itself always be on. In a non-virtual design, especially in a residential setting a non-virtual TV server may sit idle most of the day, until prime time.
Traditionally – and the way that the industry see it – you would introduce a dedicated TV server device. In an environment where already have an existing always on “24/7 devices – be it an existing server or NAS – virtualisation can allowing you to make use of your existing always-on hardware. Preventing you from having to introduce any new equipment. In essence, while your virtual TV server waits for prime time, the physical computer is doing other, more resource efficient things.
Virtualisation can therefore save you physical space (be it on the floor or in a rack). It can reduce equipment noise, reduce heat and most importantly, save power. It does so by encouraging you to spec correctly; leading to higher financial returns on equipment that you already own. So how do you create a virtual TV Server?
Virtualisation Platform
If you want to create a virtual TV server then, the platform that you choose to use will likely be the one you already have. It is easy to critique a solution and say that “you should be using something else”. Just as DVBLogic, Hauppauge and TBS have said I should use a physical device, I’ll get 50 emails telling me I should have used Proxmox, Unraid or Debian+KVM. I didn’t want to use those. I wanted to use Hyper-V.
Creating a virtual TV server is a lot easier in VMWare ESXi or KVM. Your hardware options are substantially broader due to feature maturity. For Hyper-V users, where Discrete Device Access (DDA) – Hyper-V PCIe pass through – was only introduced in 2016. The robustness of PCI Express pass-through is not yet mature and is cripplingly limiting.
Hyper-V’s issues stem from Microsoft’s design decisions. DDA follows a very robust, standards compliant implementation of the VT-d and PCI Express 3.0 specification. In 2019, most non-data centre/consumer level hardware is not manufactured to support these standards. Complicating it further, WHQL driver validation is not yet strict enough to ensure that drivers are fully compliant; and this is where most DDA related issues occur.
Hyper-V was designed to run Windows as efficiently as possible. This contrasts with their competitors, whose broader interest was to make the most efficient hypervisor platform on the market. DDA’s is a microcosm of Hyper-V’s core design limitations: Microsoft’s stated intention was to allow pass-through of select Graphic cards, GPU accelerators and NVMe controllers, not to create a robust PCIe pass-through solution. This in turn limits TV tuner hardware option.
Choosing TV Tuners
Once you understand your platform, it is important to choose your hardware accordingly.
The first discriminator will be to choose what broadcast standard you require: be it DVB-T, DVB-T2, DVB-S, DVB-S2, DVB-C, DVB-C2 or legacy Analogue.
Equally important will be matching the capabilities of your platform to the hardware device.
VMWare & KVM
VMWare and KVM derivatives offer a broader set of compatible hardware than Hyper-V. KVM is far more forgiving compared to its competitors – especially when running on non-server hardware. The chances of success are also greater if you intend to run a Linux distribution within your Virtual Machine, rather than Windows.
I have have had no luck with Hauppauge products in this regard, however there are some reports of success on-line with TBS. Comparatively, TBS offers a wider range of products, along with open source drivers. While out of reach to most users, this does offer the possibility of the community adding better support for virtualisation as platforms mature.
Reported examples of working hardware include the DVB-S2 TBS 6902 (see the comments and reviews section in the Amazon link). Despite few examples of success, getting a PCIe tuner to work reliably will remain difficult until the Tuner manufacturers migrate onto the PCI3 specification and are compelled (largely by Microsoft) to write compatible drivers.
If you wish to have a higher chance of success, with lower risk however, please follow my suggestions for Hyper-V.
Hyper-V
I was unable to get any PCIe tuner, from any manufacturer to work under Hyper-V Discrete Device Assignment (DDA). Windows VM’s would blue screen as soon as the kernel attempted to load the driver, while Linux VM’s – although stable – could not initialise the hardware device. In one set of tests, I was able to render the Hypervisor’s parent partition unusable for further testing as Hyper-V locked the hardware device and refused to release it.
After a full re-install, the situation was solved, however my testing reveals that Windows Server 2019 has not provided any improvement in using DDA with this type of legacy-bus hardware.
The solution to the problem was ultimately USB 3.0.
It is likely that your server motherboard has USB 3.0 ports on it. It is important to understand immediately that it is not possible for you to use these ports – in most cases. The embedded USB controllers on motherboards cannot usually be released to a VM by the your systems IOMMU gateway. Where they can be, it will be confusing as to which physical ports are in use, leading to difficulties in troubleshooting. Consequently, I suggest that you do not even try.
Using an inexpensive, off-brand PCIe controller from eBay . I was able to achieve a stable PCIe device pass-through with both Linux and Windows VMs under Hyper-V Server 2019. With this in place, it became possible to build a working virtualised TV Server solution.
The Software Design
Running on Hyper-V Server 2019. I installed a trial of DVBLogic’s TV Mosaic 1.0 into a Windows 10 Pro 1809 virtual machine.
TV Mosaic showing available DVB-T2 (Freeview HD) channels
DVBLogic’s trial activation system is not designed to expect to see virtual machines. Over the 3-4 months that I was experimenting, I expired trial activations for both TV Mosaic as well as its predecessor DVBLink. No matter what server, VM or physical location I tried from, I was unable to activate the trial again. If you wish to activate a trial on a VM. You will need to contact DVBLogic until such a time that they fix the issue.
The Hardware Design
Knowing that it would be necessary to replace my existing and PCIe TV Tuners with USB ones meant that I had to re-consider my design. In my original physical setup the HVR-4400 provided access to DVB-S satellite channels, with the TBS-6205 providing DVB-T2 coverage.
At the time, the only USB device that I could find to substitute the satellite tuner was the then new . The DVB-S2 device was well over £100 at the time (it has subsequently reduced considerably) and I was not willing to experiment on such a high cost tuner.
As I intended to use DVBLogic’s TV Mosaic for the project, I chose the DVB-T2 . Asking DVBLogic to support any issues would be easier if it was within their own range.
I did not want to run the TV signal down to the server rack, so chose to run the USB from the server to the signal amplifier in the attic. I purchased a good quality 5m USB 3.0 cable and a mid-cost 7 port powered hub. It was necessary to ensure that the hub used a USB type-B upstream connector to allow proper connectivity.
I already had the £12 USB 3.0 controller from a 2015 project. As will be discussed below. It is very important that the USB controller you pick has its own power connector on it. Do not rely solely on PCIe bus power.
The design was to run a single USB 3.0 5Gbps line into the attic to a powered USB 3.0 hub. The TVButler tuner would connect to the hub, and then take a short 2m coax run to the near-by signal amplifier. If the design worked. I would add additional tuners to the hub at a later time; including possibly restoring satellite connectivity.
StarTech ST93007U2C USB 3 Hub and 3x DVBLogic TVButler TV Tuners
The Final Specification
SuperMicro X11SPL-F
Intel Xeon Silver 4108
Noctua NH-U12S DX-3647, 120mm cooler for Intel Xeon LGA3647
STW USB 3.0 PCIe dual port USB 3.0 5Gbps controller
StarTech ST93007U2C 7 Port USB 3.0 Powered Hub
4x DVBLogic TVButler USB TV Tuners
LINDY Anthra Line 36744 USB 3.0 Type A to B Cable
Troubleshooting
The follow are the two main issues that I encountered when implementing the Virtual TV Server.
Single Tuner Dropouts: USB Bus Power
It was able to see the TVButler Tuner and it had a strong signal, but it would drop out after a few minutes of playback. The VM had to be rebooted to restore functionality. I removed the hub and extension cable and temporarily ran the signal down to the server rack. The issue persisted.
In my haste to minimise the hypervisors downtime. I had neglected to fit the USB 3.0 controller’s power connector. Despite using a mains powered hub, the solution was unstable. After connecting the power supply, the issue went away completely and in single tuner mode, it was stable.
Multiple Tuner Dropouts: All hubs are not created equally
After purchasing several additional TVButler tuners I setup the hub in the attic. Every 36 hours or so, I would discover that one or more of the Tuners was missing from TV Mosaic. Further investigation revealed that the tuner was missing from Windows device manager. 1 out of every 8 reboots would temporarily fix the problem.
7 out of the 8 restarts, would usually result in the driver for the bottom TV Tuner on the hub failing to load with “error 10”. Additional testing revealed that all of the tuners worked individually, as did the extension cable.
When it did work, HD channels would not play at all and SD channels would artefact as frequently as every 10 seconds.
The clue came from watching the hub while the VM rebooted. As the VM restarted, the ‘device present’ LEDs would flicker. When the reboot worked, the tuners would initialise in descending order and the LEDs remained lit. When it didn’t, the lights would enumerate randomly, flicker and, after a few seconds, the last device on the £24 RSHTECH 7 port powered hub would blink out.
Although mains powered, the flickering suggested that it didn’t have sufficient current to support the load. I swapped in a 2 port StarTech hub from my desk and with 2 tuners present and had no issues. Returning the RSHTECH it to Amazon, I ordered a – at more than double the price.
The ST93007U2C worked perfectly. All of the tuners worked properly and there were no issues at reboot.
Conclusion
As I conclude this article, the system has been in place for nearly 2 months. I have licensed TV Mosaic onto the Windows 10 VM to get around the trial issues, and it has been performing as well as I had hoped.
The Windows VM’s current uptime is 31 days, 8 minutes and at no point during the last two months have I experienced any crashes from the VM or hypervisor. Picture quality is excellent and I have artificially stress tested it to well beyond even its worst case ‘general’ use several times – with all tuners playing back HD channels while TV Mosaic transcodes the streams.
To an Intel Xeon Silver 4108, this worst-case work load is virtually irrelevant.
At idle, the server sits at around 44w, with typical non-TV load pulling 52w. Turning the TV Server VM on or off makes no difference to this figure. When TV is playing-back, this figure may rise by 8-16w. Contrast this with the old physical server, which was drawing 60-80w at all times. As a Windows 10 machine, it couldn’t function as a true server. Consequently, the Xeon Silver server would also be on anyway, taking mean idle load up to around 115w. The 71w saving (115w-44w) equates to an energy cost saving of just under £100-per year.
I spent £210.25 in total on this project, meaning that it will have paid for itself in fractionally over 2 years. If I factor in income from selling the old tuners and physical PC, I will have already broken even. So to DVBLogic, TBS and Hauppauge, all of whom queried with me the sanity of wanting to Virtualise a TV Server. You have your answer.
You can virtualise a TV Server, even on Hyper-V and, if you already have a always on “24/7” virtualisation stack. There is a good reason to do it.
