Hi ozylynx,
The UD client was developed as a simple interface for the home user. Keeping this in mind:

1. How do I monitor multiple compuers, most headless drones, in UD?
My Grid - Device Manager - Device Statistics lets you see the last time that each computer uploaded results (and then downloaded a work unit). It does not tell you which project or work unit. But this time stamp, plus the Device ID, allows a staff technician to determine this information when included in an error report. If you use BOINC and BOINCview, you can monitor all this yourself.

2. Is there a Log file in UD? This has proven essential in the past for troubleshooting.
No. BOINC has log files but not UD.

3. As I have discovered that there are problems with HPF2, (the prime reason I'm here) I have temporarily disabled that from my profile. When and how will I be notified of its resumption?
There is a lot about this that will be showing up soon in Member News and Known Issues. We are working on debugging this problem.

4. How do I identify the project number that I am working on? Needed I'm told in case of uncooperative (stuck) work.
The Device ID is shown in your client view screen. That plus the time you downloaded the unit lets the overworked grid tech access the database and figure it out. BOINC lets you read the work unit name directly and lists the Device ID at the start of the messages log.

5. Is there a way, without using BOINC that I can run multiple instances of UD on an Hyper Threading computer?
No. UD is specifically engineered to shut down multiple instances. This is because it uses a global synchronization variable that multiple instances would overwrite. Anyway, is it really smart to run multiple threads of BOINC on a hyperthreading CPU? You only have one FPU and jamming 2 giant programs into 1 L2 cache might slow things down unless you are talking about a Xeon with a really large cache.

Ask away!
Lawrence

UD is proprietary software, and is only updated by United Devices. WCG did a licensing deal with them that gets them a branded version and access to the API, but they really have very little control over the agent itself. UD's response to previous issues leaves little hope that they will add new features unless in response to a serious security bug.

BOINC is open source, so it's easy to improve and provide utilities for. Just installing and running it is no more complicated than UD, but there is a lot more complexity available if you need it, and it's infinitely easier to troubleshoot.

There's nothing stopping you from trying BOINC out and removing it later. It will even work in parallel with UD, but that's not a recommended configuration.

The only score that counts is the one shown on the website.

Hello ozylynx,

I have multiple computers (18) running the UD client. All except two of them have no screen, keyboard or mouse. They are all Windows XP-Pro machines and I manage all of them from my main PC using the in-built Remote Desktop facility. Provided you allocate fixed IP addresses to each machine it is easy to take remote control of whichever PC you want and see what is happening to the UD agent.

In my own case I find it sufficient every two or three days to click on "Device Installations" under My Statistics in My Grid on the WCG website to see when each agent last returned a result. Looking at the bottom of the list I will find any agent that has been running longest without returning a result. If it looks as if any agent hasn't returned a result for a suspiciously long time I then log on to the appropriate PC and investigate.

I usually log onto each machine routinely every week or so to confirm all is running well and to see if any Windows Updates or other housekeeping needs doing.

I wonder if this kind of regime might be a manageable solution for you?

Kind regards,
David.

I am running BOINC and have tested several projects. I have used the UD interface also.

If you insist on staying with the UD interface there is an add-on product called UDMonitor. It is intended for queueing up multiple devices to process offline before calling in again.

Conversely, you can just stay with one device if you want and still run UDMonitor. The log in UDMonitor does a nice job of monitoring your checkpoints, with the timestamp for each and the time between them. In addition, the flask indicator changes from blue...to yellow....to red.....and back to blue when it just passed a checkpoint.