Folding Forum

bruce wrote:Did you find the specific log that covered a few hours before 2012-12-09 13:08:42 UTC (aka 21:08:42 PST)?

I'll confess to causing that one. Windows update ran and did a GUI reset and it fragged that one. But HFM.NET has no record of it in the WU History database.

I am checking into the pausing and core updates now. The slot would have taken 1 core update, but not 4. It has been steady on 807x since it received it's first one.

I wish it would count 404s from the servers as errors instead of pauses (which I know is in the works).

P5-133XL wrote:Do you pause v7? HFM.net when it examines the folding logs it counts pauses as failures when they obviously are not.

HFM does not examine the logs. It gets all of its information through the telnet interface.

codysluder wrote:HFM does not examine the logs. It gets all of its information through the telnet interface.

Do you know where the "failed" data persists?

I don't see it in the database and I don't see it in the logs. Multiple restarts to roll the current logs does not shake the "false" failed units. I do see some *.dat files, but not sure what is in them or if they are safe to remove for eliminating some bad history.

Right now, I am only seeing the problem on V7.x.x clients. I even see a fail on a GPU that is set not to fold, it is set not to start.

Try restarting HFM.net, that normally resets its failed counts.

PinHead wrote:

codysluder wrote:HFM does not examine the logs. It gets all of its information through the telnet interface.

Do you know where the "failed" data persists?

I don't see it in the database and I don't see it in the logs. Multiple restarts to roll the current logs does not shake the "false" failed units. I do see some *.dat files, but not sure what is in them or if they are safe to remove for eliminating some bad history.

Right now, I am only seeing the problem on V7.x.x clients. I even see a fail on a GPU that is set not to fold, it is set not to start.

See this for information on why v7 does this.
https://groups.google.com/forum/?fromgr ... BiWaVtfMiY

Has anyone noticed a significant increase in HFM CPU usage after the monitored v7 clients and HFM has been running for a week or more ?

toTOW wrote:Has anyone noticed a significant increase in HFM CPU usage after the monitored v7 clients and HFM has been running for a week or more ?

I know there were a few problems in an earlier version but i haven't seen anything amiss with 0.9.1, but then i don't leave hfm up for more than a day or so at a time

I'm using HFM 0.9.1 rev 595 64 bits on a Windows 7 machine.

It's monitoring 3 clients : one v6 client using HTTP and two v7 clients. One v7 client has one SMP slot, and the other SMP client has one SMP and two GPU slots. It is also exporting XML stat for my monitoring page : http://fahmon.fleucorp.fr

When everything is started, my i7 running HFM scores 18k PPD, but after a week, it's down to 12k PPD. I notice that HFM uses a full core for more and more time on each update.

I miss the v6 clients and FahMon

I too noticed full core usage when I upgraded to the latest version. I decided to do a clean install and the problem went away. I did nuke all data but it's value was questionable anyway with all the changes that were happening.

That was last september and it's behaved well since.

My original post about the problem.

Well ... it's a clean install from the end of december ... so I don't know what could have went wrong with it so soon ...

Do you have the facilities to do an IP trace? If so, are there there anything noteworthy changes in the patterns of the telnet messages between HFM and FAHClient?

i'm only running 2 v6 clients at the moment but i'll leave my daily driver on for this week running hfm and see if i notice anything strange. I'll grab some screenies as well

toTOW wrote:I'm using HFM 0.9.1 rev 595 64 bits on a Windows 7 machine.

It's monitoring 3 clients : one v6 client using HTTP and two v7 clients. One v7 client has one SMP slot, and the other SMP client has one SMP and two GPU slots. It is also exporting XML stat for my monitoring page : http://fahmon.fleucorp.fr

When everything is started, my i7 running HFM scores 18k PPD, but after a week, it's down to 12k PPD. I notice that HFM uses a full core for more and more time on each update.

I miss the v6 clients and FahMon

I have a similar setup with 2 V6 clients and 2 V7 clients on Windows 7 64 bit that uploads stats to a web page, I haven't noticed CPU usage per say. But sometimes it seems to cause a datastorm on the network possibly because the 2 V7 clients that are running also have one to two detected possible clients that are not running. Next time it storms, I'll check cpu usage; but the idle V7 clients which I can't seem to remove ( because it removes all clients on V7 and not just the idle ones ) might be causing my issues as they are never going to respond to the polling.

When it storms, I have to shutdown HFM.NET to stop it. But under normal conditions I have not noticed CPU usage.

I've just downloaded and set up FAH v7 and updated HFM to latest version but seem to have losty ability for HFM to look at historic projects using F8 key - works fine on older clients and I can see current projects but is there any way of looking back at history using HFM or is it a matter of delving into the V7 log.
Could not find anything about this in forums and hope I'm not wasting your time. Thanks for any help and / or pointers
Shunter