Questions and Answers :
Unix/Linux :
Multiple computers setup?
Message board moderation
Author | Message |
---|---|
ScottieD369 Send message Joined: 8 Jun 15 Posts: 26 Credit: 264,597 RAC: 2 |
I have a small mini Raspberry Pi cluster setup. (PoE) They all are running 4 each tasks each. When I login to account under my computers, not "all" of them are listed? I am wondering why? I am guessing that maybe they are working together on the same projects? Or the ones missing aren't able to connect? But yet as I said ALL of them are working and doing tasks and reporting them. So any ideas what I can do? If else I'll just leave them as is since it does seem to be working?...... PS if I fix this and they all get to show? Will I get a higher RAC #? Thanks |
rob smith Send message Joined: 7 Mar 03 Posts: 22448 Credit: 416,307,556 RAC: 380 |
Simplest first... Each of the four cores on each of your RPi will be working on its own task as SETI is not capable of using more than one "cpu" core. My first thought was that they were very new (a few hours old), but the three I can see date back to early December last year. So, next thought is that you used two (or more) different email addresses, with two using the address you've used to sign in here, and the others with a different one. Obviously the more you have working on a single account the more credit you will get. If they are RPi 4 you will probably get between 600 and 800 each - a bit more if they are well cooled, a bit less if they are running hot. Bob Smith Member of Seti PIPPS (Pluto is a Planet Protest Society) Somewhere in the (un)known Universe? |
ScottieD369 Send message Joined: 8 Jun 15 Posts: 26 Credit: 264,597 RAC: 2 |
No they are all on the same Email account. I have no clue to why this is happening? They are all linked to same account but but all is listed. The ones I have are currently 3B+ not 4. |
rob smith Send message Joined: 7 Mar 03 Posts: 22448 Credit: 416,307,556 RAC: 380 |
Credit will be somewhere between half and two-thirds of a RPi4. Thinking what diagnostics to run and work out what is going on...... (I might be tomorrow before my brain wakes up again - 4am (UTC) is a long way back in history) Bob Smith Member of Seti PIPPS (Pluto is a Planet Protest Society) Somewhere in the (un)known Universe? |
ScottieD369 Send message Joined: 8 Jun 15 Posts: 26 Credit: 264,597 RAC: 2 |
Then yes everything "seems" to be working AND reporting just not displayed as a linked computer? They all have the same computer names and logins. |
rob smith Send message Joined: 7 Mar 03 Posts: 22448 Credit: 416,307,556 RAC: 380 |
Each should have a unique ID number (only you can see the names) - can you check these numbers and post them here (might be a bit of an issue if you can't see them on the website). Bob Smith Member of Seti PIPPS (Pluto is a Planet Protest Society) Somewhere in the (un)known Universe? |
ScottieD369 Send message Joined: 8 Jun 15 Posts: 26 Credit: 264,597 RAC: 2 |
Nope I am not able to see them numbers. By looking at the ones that are there it seems like 5 are all linked together as 1? |
Mr. Kevvy Send message Joined: 15 May 99 Posts: 3797 Credit: 1,114,826,392 RAC: 3,319 |
|
ScottieD369 Send message Joined: 8 Jun 15 Posts: 26 Credit: 264,597 RAC: 2 |
Well either way it is still not correct. It is still split up weird? |
Mr. Kevvy Send message Joined: 15 May 99 Posts: 3797 Credit: 1,114,826,392 RAC: 3,319 |
Running SETI@Home or any other BOINC project on a cluster probably is not very efficient. Cluster computing is for large single jobs that have not already been split into smaller ones for parallel processing. BOINC's work units already have been, so the overhead of having to divvy up the calculations is redundant and will detract from the output. BOINC already will use all the CPU on each node so the master node won't be able to get any more work out of the compute nodes than they already would do with just BOINC running on each as single machines. |
rob smith Send message Joined: 7 Mar 03 Posts: 22448 Credit: 416,307,556 RAC: 380 |
Mr Kevvy's comment would apply if you were running something like a Beowulf cluster, but in that case I would expect to see a single computer with a lot more than 4 CPUs on your list of computers, but there are actually three(!?) showing, each with four cores: ID: 8881835 ID: 8878096 ID: 8864428 So, let's take a step back.... How many RPi do you actually have? (If it is five then there may be two "missing") Can you connect a screen to each of them, in turn, and using the "advanced view" see what the stats are for each one? (I'm hoping here that you are using the BOINC GUI as that makes life a bit easier) - on the "projects" tab the "work done" and "average work done should be more or less the same. Right - I've found where to locate the computer id number - be warned this is not a safe or simple place to be fiddling around... In the BOINC data directory, you should find a foler "slots", and within that a set of folders number 1 to x - in each of those there should be a file called "init_data.xml", in the first few lines of this file you will find one like: "<hostid>12345678</hostid>" - that number is the unique id for every computer that has ever attached to SETI@Home. While you are looking in the file check that the user id number is the same (that is your user identity). DO NOT try and edit the file as doing so could well result in some rather undesirable effects..... Bob Smith Member of Seti PIPPS (Pluto is a Planet Protest Society) Somewhere in the (un)known Universe? |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14674 Credit: 200,643,578 RAC: 874 |
Right - I've found where to locate the computer id number - be warned this is not a safe or simple place to be fiddling around...Much easier and safer way. Use BOINC Manager, as Rob says. In Advanced view, Projects tab, click on (highlight) the SETI@Home line. In the left hand column, click 'Properties' (bottom button, upper group). Mine says (and this is a copy'n'paste) Computer ID 8313133 Most of the other useful fields are visible too. While you're on each machine, set up Controlling BOINC remotely - then you won't have to keep switching those cables over - sit at a central point, and connect to each machine electronically to see what's going on inside it. |
ScottieD369 Send message Joined: 8 Jun 15 Posts: 26 Credit: 264,597 RAC: 2 |
I actually have (7) pi's connected at the moment. I'll have to get back to you on the #'s. |
rob smith Send message Joined: 7 Mar 03 Posts: 22448 Credit: 416,307,556 RAC: 380 |
Thanks Richard. (And I keep forgetting to set remote control on.....) Scottie - when I last looked there were three listed, now four - they must be shy Bob Smith Member of Seti PIPPS (Pluto is a Planet Protest Society) Somewhere in the (un)known Universe? |
ScottieD369 Send message Joined: 8 Jun 15 Posts: 26 Credit: 264,597 RAC: 2 |
That's weird but here are the #'s. (Mental note: Bottom Up) Raspberry Pi' #'s 8864428 Below here is the mini cluster that was setup recently. With update. 8881835 8881835 8881835 8881835 8881835 8878096 Laptop #'s 8786501 So yeah there are (5) Pi's that have the same ID. Any clue to how this happen or to fix? Yes I have setup all the same way and same Email. Those 5 were all setup in the same day within the same hour all back to back. Then on the last one is different was setup at a later date and added into the collective same setup. I know the laptop is irrelevant but I thought I'd add in all my #'s that I see and I gave up on my mobile app. The single Pi at top is a Pi 2B+ setup years ago, and the mini cluster are all 3B+ and setup a few months ago. Also another side note is that the 5 are all connected to the same server #. All I did was Install BOINC login the SETI @HOME and ran. |
Kissagogo27 Send message Joined: 6 Nov 99 Posts: 716 Credit: 8,032,827 RAC: 62 |
2K rac isn't bad for four pi ^^ Server seems to see them as only one ... |
ScottieD369 Send message Joined: 8 Jun 15 Posts: 26 Credit: 264,597 RAC: 2 |
So I am wondering what is wrong here? AS I said everything is working. If I have to just deal with it, it is ok. But if there is anything I can do to fix this I am willing to try. |
rob smith Send message Joined: 7 Mar 03 Posts: 22448 Credit: 416,307,556 RAC: 380 |
They are all working, the RAC is about right for four RPi running. But why the same ID? - If they are running as a "proper" cluster then that would explain things, but you would have had to set up the cluster first and would probably had to install some clustering software onto them all. What is strange is that they report as only having four cores active, not the 16 I would expect, but that could be down to the clustering software. Another thought, did you get one of the batch of four working then clone its memory card onto the other four, given the low performance of the RPI that might explain what is going on, but I'm not sure.... Bob Smith Member of Seti PIPPS (Pluto is a Planet Protest Society) Somewhere in the (un)known Universe? |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14674 Credit: 200,643,578 RAC: 874 |
If each Pi is running independently, and contacting the SETI server independently, then the server will get told the details of one single Pi on each contact - 4 cores, and so on. But the server is supposed to have checks against duplicate ID numbers, and should have spotted this and given each Pi a different number after the first round of contacts - even if they were set up by cloning an SD card. The strange thing for me is that host 8881835 is recorded as only having contacted the server 6 times since 10 Jan 2020, despite having 124 tasks. |
rob smith Send message Joined: 7 Mar 03 Posts: 22448 Credit: 416,307,556 RAC: 380 |
One check that may help us see how big the wood is, and thus find the tree we want: If you can see the access record on your router it may give a hint as to what is going on - I would expect, if the RPI are behaving as a cluster, there to be a single MAC address, but if they are independent then then there would be several MAC and IP addresses, one for each. This is certainly a strange one. Bob Smith Member of Seti PIPPS (Pluto is a Planet Protest Society) Somewhere in the (un)known Universe? |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.