FahMon

Whitestar · 11 Jul 2007 at 13:18

Is it me or does FahMon not like monitoring networked clients?

I wanted to monitor my whole farm from fahmon, but even with just 20 clients it seems to crash on a daily basis.

verbal · 11 Jul 2007 at 13:20

Hi Whitestar, are you running it in Windows or Linux?

shadowscotland · 11 Jul 2007 at 14:49

I never got that to work (with our 'exotic' proxy)
If you can't get it working, I use an alturnative web based utility hosted by another team that may be of help (the programmer is cool with us using it)

Really usefull for me monitor work pc's when I'm at home

My page is here the yellow warning are due to pc's that are not on 'enough'. you even get full histories of each pc by clicking on them and it works with linux too

Can beat Fahmon for looks but has lot's of usefull data to view.

Whitestar · 11 Jul 2007 at 16:58

@ Verbal, I'm using it in windows.

@ SS, I remember you saying about that site before. Would they be ok with me adding all my clients?

shadowscotland · 11 Jul 2007 at 17:38

I've asked and they are cool with us all using it (but no real point if all your eggs are in one basket)

One of their team has 27 pc's on his account and there's a way of grouping them so you could have departments at the top level with separate pc in that group (from memory it's like sub-teams with the [])

All pc's must have a unique number/code within your user name, that very important. I use our telephone extension numbers as the ID.

How many clients do you run? is 173 correct?

If it's that many I'll ask egon (programmer) and get back to you.

Whitestar · 11 Jul 2007 at 18:54

Yeah, thats right for now, but it will be going up in the near future.

shadowscotland · 11 Jul 2007 at 19:05

you do realise that only 35ddp per machine in real terms.

- That's well, carp

(aka same as my p4's here which are only one 9-6 now

)

Edit - just check with egon and his exact words were - bring it on! :cool:

You got to love americans

Instructions here
and that going to take you ages to set 170+ machines up

Edit 2 - just told him it might be 300

Reply below

Egon said:
Re: Fahmonitor
From Egon | 07/11/2007 12:09:26 PM PDT new

Excellent!

I think our high was around 250 at one time. Klutz Dohanger was our primary person with close to 60 registered... but he’s fallen off considerably— not in production, but in actually monitoring his systems.

It shouldn’t be an issue. There’s not much data going across, actually, and if they’re only coming across every half an hour or so...

One thing you might want to do, though, if you don’t mind. Take groups of them and offset them by 5 minutes or so, so they’re not /all/ hitting at the top and bottom of the hour!

shadowscotland · 12 Jul 2007 at 09:40

Egon asked me to say hold off for a day - as he's trying to sort my 'wrong time zone' issue - been fault finding all morning so should have someting sorted some

Whitestar · 12 Jul 2007 at 11:06

Cool. It will take me a while to set up all the machines and i'll have them set to upload once a day.

rich99million · 12 Jul 2007 at 12:57

Just wondering if you're using the latest version of FahMon? The previous version could take ages to contact network machines but I thought it was all sorted :confused:

uncle_fungus · 12 Jul 2007 at 13:59

I'd like to know how it crashes too, and to get a copy of messages.log if possible. That might help me work out how/why FahMon crashes.

Whitestar · 12 Jul 2007 at 14:35

Well, I have a feeling it was crashing because I had the automatic reload time too low. I had it set at half an hour, then I set it to an hour, but it still kept crashing. I've now set it to two hours and it hasn't crashed yet.

I think the problem is that its trying to reload clients that it still hadn't finished trying to reload from the last time. I dunno. :confused:

uncle_fungus · 12 Jul 2007 at 19:20

Whitestar said:
Well, I have a feeling it was crashing because I had the automatic reload time too low. I had it set at half an hour, then I set it to an hour, but it still kept crashing. I've now set it to two hours and it hasn't crashed yet.

I think the problem is that its trying to reload clients that it still hadn't finished trying to reload from the last time. I dunno.

Well unless there something really screwy going on in Windows, it shouldn't take that long to update clients. However, if the FAHlog.txt file is really huge, try turning "show fahlog" off and see if it helps the speed.

There may be some as yet unresolved bug with the update routines, which are being exacerbated by your many clients and low update interval. If that's the case, setting a larger update interval would only increase the mtbf. It would be useful to know, either way.

SiriusB · 12 Jul 2007 at 20:12

Can FahMon handle 170+ machines at the same time?

Id be tempted to have multiple instances of FahMon running, say one for each room [assuming all the PCs aren't in one big computer lab

]

uncle_fungus · 12 Jul 2007 at 20:18

SiriusB said:
Can FahMon handle 170+ machines at the same time?

Id be tempted to have multiple instances of FahMon running, say one for each room [assuming all the PCs aren't in one big computer lab ]

Well according to FLECOM, it could handle at least 700 clients, although he was running a version prior to when I took over development. I *hope* I haven't broken that stability/capability. http://forum.folding-community.org/fpost164859.html#164859 If I have, then that will jump straight to the top of my development list.

Whitestar · 13 Jul 2007 at 16:36

I havn't got all 170 on yet, Ive only got 20. Unfortunatly my netowrk has thrown a wobbler, so fahmon won't connect to any remote clients. Although, saying that, it hasnt crashed.

shadowscotland · 18 Jul 2007 at 17:25

the new script is here if you need it whitestar

http://vulcan.pve-inc.com/fah/FAHUpload-Win.zip