I've noticed during this period of short 'good queues' and long 'bad queues', there is no indicator on the status board that reflects this fact. I'd like to see some sort of 'files to delete' indicator that would theoretically be going down while the validation queues are growing. I'm sure one or two other indicators would be really neat as well. Maybe someone can make a suggestion?

---

May this Farce be with You

<blockquote>Maybe someone can make a suggestion?</blockquote>
Didn't you just do that? :)
A for this idea.

Matt Lebofsky, if you see this -- a lot of people would like some more info on the Server Status Page.

BTW, thanks for the extensive update on Technical News. It feels good to be informed! :)

I think with all the added server roles and additional servers the list of servers could be rearranged better, perhaps in a table with validators in one column, assimilators in another etc.

Keep up the good work guys, good to have the information in tech news.

---

Raithmir's SPARC64/UltraSPARC Linux Builds
http://www.kulthea.net/boinc/

Another nice feature would be the return time of the oldest result still in the validation Q

That way we could see where it was up to and that it was actually processing results as all we can currently see is the validation Q slowly creeping upwards

---

Part of Teamseti
For SetiBoinc status graphs visit Teamseti status graphs

Another nice feature would be the return time of the oldest result still in the validation Q

That way we could see where it was up to and that it was actually processing results as all we can currently see is the validation Q slowly creeping upwards

Good idea!

---

Raithmir's SPARC64/UltraSPARC Linux Builds
http://www.kulthea.net/boinc/

A few days ago I tossed together a little script that pulled bits of data from the status page and saved them. Mostly just for my own use to see what the bottom line trends were. By no means a definitive picture of how things are working, but it lets me see what the queues were like over a period of time. For anyone interested, the output is here.

A few days ago I tossed together a little script that pulled bits of data from the status page and saved them. Mostly just for my own use to see what the bottom line trends were. By no means a definitive picture of how things are working, but it lets me see what the queues were like over a period of time. For anyone interested, the output is here.

Nice thats handy.

It has confirmed what I believed that there is a general upwards trend to the validation Q.

I am sure that the validators are processing units, however as per my comment above it would be nice to get a little bit more info on the validator Q to see where it is standing atm.

---

Part of Teamseti
For SetiBoinc status graphs visit Teamseti status graphs

A few days ago I tossed together a little script that pulled bits of data from the status page and saved them.

Nice!
Would you mind sharing your script?
Got a similar idea, didn't start writing it yet.
When I saw your post, I wondered why I should reinvent the wheel.

Nice!
Would you mind sharing your script?
Got a similar idea, didn't start writing it yet.
When I saw your post, I wondered why I should reinvent the wheel.

The original script is a long, drawn out bunch of Perl code with the data extracted getting put into a mysql database, etc, etc. Obviously sort of like swatting a mosquito with a nuclear bomb. In trying to apply the KISS principal to it I punched in this bash script that accomplishes approximately the same out put without the need for a database and all the other excess baggage. I'm sure there's TONS of room for improvement, and it assumes a Linux environment with the usual penguin-like utilities, but it's a quick and dirty stab at making it a bit less bulky. Feel free to view it, download it, paint it purple, or do anything else ( morally acceptable) with it that you wish to. :)

Bash Code

A few days ago I tossed together a little script that pulled bits of data from the status page and saved them. Mostly just for my own use to see what the bottom line trends were. By no means a definitive picture of how things are working, but it lets me see what the queues were like over a period of time. For anyone interested, the output is here.

Now, it'd be really cool if you could dump those numbers into MRTG or something like that.

---

Very nice -- seeing trend data is a very good thing!

A few days ago I tossed together a little script that pulled bits of data from the status page and saved them. Mostly just for my own use to see what the bottom line trends were. By no means a definitive picture of how things are working, but it lets me see what the queues were like over a period of time. For anyone interested, the output is here.

---

Now, it'd be really cool if you could dump those numbers into MRTG or something like that.

hmmm... never thought about that. We've got several scripts that read the lmsensors stuff, plop it into an RRD database, and feed it all into gnuplot to make graphs for system temperatures. Bet I could modify one of those rascals to make some eye candy out of all those numbers. Been so long ago since I did all that stuff I'll have to snoop through it and see if I remember how it all works.

http://teamseti.com/status/status.phtml

Just set up an rrd graph set based on the 4 metrics shown on the status page.

If there are any rrdgraph gurus out there who fancy giving me a hand to nicen it up a bit then just let me know.

Cheers

---

Part of Teamseti
For SetiBoinc status graphs visit Teamseti status graphs

I'm sure there's TONS of room for improvement, and it assumes a Linux environment with the usual penguin-like utilities, but it's a quick and dirty stab at making it a bit less bulky. Feel free to view it, download it, paint it purple, or do anything else ( morally acceptable) with it that you wish to. :)

Thanks for sharing it. :)
It does pretty much fit my needs already, and it's well documented.
Good stuff ;)

Now, it'd be really cool if you could dump those numbers into MRTG or something like that.

hmmm... never thought about that. We've got several scripts that read the lmsensors stuff, plop it into an RRD database, and feed it all into gnuplot to make graphs for system temperatures. Bet I could modify one of those rascals to make some eye candy out of all those numbers. Been so long ago since I did all that stuff I'll have to snoop through it and see if I remember how it all works.

I've spent no time with RRD, so I can't comment a whole lot.

You can feed MRTG a file with four numbers in it, and basically it'll swallow those and dump them into its own database -- and since it can use RRD that might be enough....

---

You can feed MRTG a file with four numbers in it, and basically it'll swallow those and dump them into its own database -- and since it can use RRD that might be enough....

Yes its doing basically that already using rrdtool

I was thinking about adding extra data to the graphs like a text line showing the current value, min max and average type of thing.

Unfortunatly I cannot find any well explained examples of the gprint param and selecting info to use in the lines


I am currently using


rrdtool graph /var/www/teamseti.com/status/seti_validate.gif --title="Waiting for validation" \
DEF:validate=/var/lib/rrdcollect/seti.rrd:validate:AVERAGE \
AREA:validate#00a000:"Validation Qc"


But it seems to plot blank space at the end of the graph for a while then suddenly backfill it in after a while, even though there is data for the blank space in the rrdb

Is this to do with using the AVERAGE CF maybe?

---

Part of Teamseti
For SetiBoinc status graphs visit Teamseti status graphs

[But it seems to plot blank space at the end of the graph for a while then suddenly backfill it in after a while, even though there is data for the blank space in the rrdb
Is this to do with using the AVERAGE CF maybe?

It's been a while since I set rrd up for system temps and various other measurments, and I haven't had a chance to look back over the scripts, but I do recall similar behavior when first starting them up. We had dropouts and some blank data initially but as it ran for a few hours it got itself syncronized and has had a steady graphed output since then.
By the way, excellent job thus far, graphs look great!

Well I've found out the cause of the blank space.

Made a typo creating the rrd and specified 3000 instead of 300 seconds as the period to average

Ive restarted the graphs now and hopefully they will start to look a bit more accurate and a bit less like straight lines

---
EDIT

Yes thats starting to look better.

Just gotta wait for a few houre for it to show the trend again as I had to wipe the RRA due to the data now wanting to restore into the new format rra

---

Part of Teamseti
For SetiBoinc status graphs visit Teamseti status graphs

By the way, excellent job thus far, graphs look great!

Thanks a lot

Hopefully they will be useful to someone (just gotta wait for some more data in them again ) :)

---

Part of Teamseti
For SetiBoinc status graphs visit Teamseti status graphs