SETI@home | Backing off 2 hr 45 min 50 sec on download

Message boards : Number crunching : SETI@home | Backing off 2 hr 45 min 50 sec on download
Message board moderation

To post messages, you must log in.

Previous · 1 · 2 · 3

AuthorMessage
Profile Wiggo
Avatar

Send message
Joined: 24 Jan 00
Posts: 35004
Credit: 261,360,520
RAC: 489
Australia
Message 1206905 - Posted: 17 Mar 2012, 10:16:42 UTC - in response to Message 1206902.  
Last modified: 17 Mar 2012, 10:17:09 UTC

I have SIV installed, I get 20 vlars sent to me, and every single one came in at 20-30Kb/s without a hang or stall during d/l..

I would bet several beers sooner of later SIV will need to kick in.

I find a lot depends on which boinc2.ssl.berkeley.edu you get as 208.68.240.13 seems a lot better than 208.68.240.18 in terms of transfer speed. I found a few posts saying the same and wonder if this is an issue that needs to be addressed.

This connection problem has been going on for some time now but it's being caused by hardware that is not SETI though when AP's are being handed out and saturating the connection either .13 or .18 (or both) start dropping packets though using a good working proxy will get most people around this bit of flakey hardware.

The lab guys have been trying to locate this bit of hardware for some months now.

Cheers.
ID: 1206905 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14654
Credit: 200,643,578
RAC: 874
United Kingdom
Message 1206909 - Posted: 17 Mar 2012, 10:35:04 UTC - in response to Message 1206905.  

I have SIV installed, I get 20 vlars sent to me, and every single one came in at 20-30Kb/s without a hang or stall during d/l..

I would bet several beers sooner of later SIV will need to kick in.

I find a lot depends on which boinc2.ssl.berkeley.edu you get as 208.68.240.13 seems a lot better than 208.68.240.18 in terms of transfer speed. I found a few posts saying the same and wonder if this is an issue that needs to be addressed.

This connection problem has been going on for some time now but it's being caused by hardware that is not SETI though when AP's are being handed out and saturating the connection either .13 or .18 (or both) start dropping packets though using a good working proxy will get most people around this bit of flakey hardware.

The lab guys have been trying to locate this bit of hardware for some months now.

Cheers.

Yes, you're right - I should have mentioned that too. From Technical News

Speaking of network competition - yes, we're aware that we are dropping all kinds of connections during uploads/downloads. This isn't because of our router (which was definitely the problem over the summer before we added RAM to it), but somewhere else further up the pipeline. Still figuring this out, but it's certainly load related.
ID: 1206909 · Report as offensive
Profile cliff
Avatar

Send message
Joined: 16 Dec 07
Posts: 625
Credit: 3,590,440
RAC: 0
United Kingdom
Message 1206927 - Posted: 17 Mar 2012, 11:53:17 UTC - in response to Message 1206909.  

Hi Richard,
Some time ago I used visual route to trace between my location and s@h, it showed an apparent problem prior to Berkeley, at what I assume was the pipeline provider, and then an unreachable node at Berkeley. Possibly that was it hitting a firewall.

But maybe all the problems are'nt internal to the s@h site.And one set of problems is masking another..

Regards,
Cliff,
Been there, Done that, Still no damm T shirt!
ID: 1206927 · Report as offensive
Profile cliff
Avatar

Send message
Joined: 16 Dec 07
Posts: 625
Credit: 3,590,440
RAC: 0
United Kingdom
Message 1206929 - Posted: 17 Mar 2012, 11:58:38 UTC - in response to Message 1206902.  

Hi Ray,
Well just got a couple more groups of WU sent, and not one hung..
But I'm pretty sure that the problem will arise again:-)

And I'll not take any bets on it:-)

As I understand it the issue is being looked at, however I think its more than one issue, and problably there are others being masked behind whatever is at 1st found..

I'm pretty sure from past experience that there is always another nasty just waiting to take a bite out of you if you dont stay alert.

Regards,
Cliff,
Been there, Done that, Still no damm T shirt!
ID: 1206929 · Report as offensive
Profile cliff
Avatar

Send message
Joined: 16 Dec 07
Posts: 625
Credit: 3,590,440
RAC: 0
United Kingdom
Message 1206933 - Posted: 17 Mar 2012, 12:25:24 UTC
Last modified: 17 Mar 2012, 12:27:18 UTC

Who did what?
I just got another load of vlars, d/l rate in excess of 91Kb/s, no stalls, no hangs, just wham bang and down.17 cuda & 47 cpu..

Is this a record for uninterupted downloads?


I'm going to have to put a postit note on the wall to comemorate the event:-)

Cheers,
Cliff,
Been there, Done that, Still no damm T shirt!
ID: 1206933 · Report as offensive
Profile Wiggo
Avatar

Send message
Joined: 24 Jan 00
Posts: 35004
Credit: 261,360,520
RAC: 489
Australia
Message 1206934 - Posted: 17 Mar 2012, 12:28:58 UTC - in response to Message 1206927.  

Hi Richard,
Some time ago I used visual route to trace between my location and s@h, it showed an apparent problem prior to Berkeley, at what I assume was the pipeline provider, and then an unreachable node at Berkeley. Possibly that was it hitting a firewall.

But maybe all the problems are'nt internal to the s@h site.And one set of problems is masking another..

Regards,

Cliff, this is a problem that many off us here over the last several months have been trying to find and it has nothing to do with SETI as I've already stated ( and Matt has already posted about, which is the quote that Richard posted) but if you know a way to do tracert through a good working proxy and without at the same time while AP's are are causing the overload of this piece of hardware then do so as this will be the only way to find it but none of us here have been able to do so.

Cheers.
ID: 1206934 · Report as offensive
Profile red-ray
Avatar

Send message
Joined: 24 Jun 99
Posts: 308
Credit: 9,029,848
RAC: 0
United Kingdom
Message 1206955 - Posted: 17 Mar 2012, 13:49:00 UTC - in response to Message 1206933.  
Last modified: 17 Mar 2012, 14:01:57 UTC

Who did what?
I just got another load of vlars, d/l rate in excess of 91Kb/s, no stalls, no hangs, just wham bang and down.17 cuda & 47 cpu..

Is this a record for uninterupted downloads?

I'm going to have to put a postit note on the wall to comemorate the event:-)

This may be down to Cur: 81.58 Mbits/sec on the Crickets (http://fragment1.berkeley.edu/newcricket/grapher.cgi?target=/router-interfaces/inr-250/gigabitethernet2_3&ranges=d%3Aw&view=Octets) which are usually at 93 Mbits/sec or more.

Maybe I should get the SIV [BOINC Status] page to show the current Cricket D/L rate. What do others think?
ID: 1206955 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14654
Credit: 200,643,578
RAC: 874
United Kingdom
Message 1206957 - Posted: 17 Mar 2012, 13:52:41 UTC

If you've looked at the Cricket graph in the last six hours or so, you'll see that the download traffic has started to tail off - the AP splitters aren't running at the moment.

We normally see good download speeds when the pipe isn't fully saturated: as Matt put it, "it's certainly load related". So don't start counting your chickens yet.
ID: 1206957 · Report as offensive
Profile HAL9000
Volunteer tester
Avatar

Send message
Joined: 11 Sep 99
Posts: 6534
Credit: 196,805,888
RAC: 57
United States
Message 1207028 - Posted: 17 Mar 2012, 16:30:17 UTC - in response to Message 1206897.  

As issues go this one won’t cause me to shed many tears, especially so given SIV reported "HKEY_LOCAL_MACHINE\SOFTWARE\Space Sciences Laboratory, U.C. Berkeley\BOINC Setup missing". I did not know BOINC would run if the registry was not setup and given you set all 3 entries to D:\BOINC\ I expect this will only work when both the BOINC programs and client data reside in the same rather than using 2 separate folders. Is this the case?

I am none too sure how I should make SIV work without the registry settings as it does not know where to find BOINC. I guess it could check for BOINC.exe being in the current folder and get the SIV start-up code to create the “missing” registry information. With your setup this would work if you either put all the SIV files in D:\BOINC\ or started SIV with D:\BOINC\ as the “Start in” location.

Would you be happy for SIV to create the registry entries and would the proposed solution be good enough?

If you don’t wish SIV to create the registry keys and values then I expect I could make it work without doing this. It would just be a slightly more pervasive change.

Yeah most issues won't make developers cry unfortunately. :(

BOINC doesn't require and registry entries to exist when you run it. It will make the entry "HKCU\Software\Space Sciences Laboratory, U.C. Berkeley\BOINC Manager", but that just stores the GUI settings for boincmgr.exe I believe. I would have to double check to see if that entry is made on machines where I don't run the manager, which is about 97% of the boxes I use.

When you start BOINC in this manor as there are no settings to tell it where the data directory is located. So it assumes its running directory is the data directory. Probably because that is how it originally worked many moons ago. This is great for how I run BOINC on my machines. Less hassle than having to worry about making registry entries whenever I ghost a machine.

I personally don't like having the entries pointing to the locations as it is an issue when I run other instances of BOINC. If I have say D:\BOINC\ and D:\BOINC2\ on the same machines having entries for either location would cause the other not to function correctly. Also if I need to test something out I will have another instance in another location for whatever I'm doing. While the main one is also running.

I tend to think the best option is look in the registry for the entry and then require input if it can't be found. The rub in that is this isn't dedicated BOINC type software. So you hide the function if not found. Perhaps something like -BOINC_PATH= as a start parameter? As an override to looking in the registry.
SETI@home classic workunits: 93,865 CPU time: 863,447 hours
Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[
ID: 1207028 · Report as offensive
Josef W. Segur
Volunteer developer
Volunteer tester

Send message
Joined: 30 Oct 99
Posts: 4504
Credit: 1,414,761
RAC: 0
United States
Message 1207075 - Posted: 17 Mar 2012, 17:32:40 UTC - in response to Message 1206955.  

red-ray wrote:
...
Maybe I should get the SIV [BOINC Status] page to show the current Cricket D/L rate. What do others think?

Scraping data from an HTML page is often frowned upon. Having potentially thousands of hosts running SIV downloading that status just might make it impossible for S@H staff to ever get an agreement with the campus network people allowing more up/download bandwidth by using part of the 1 Gbps link to SSL.

I don't mean to be alarmist, most probably if the question were asked the campus people would say that in the big picture of how much traffic they're handling it would do no harm. But without the question, it could be used as a point against the project.
                                                                   Joe
ID: 1207075 · Report as offensive
Profile red-ray
Avatar

Send message
Joined: 24 Jun 99
Posts: 308
Credit: 9,029,848
RAC: 0
United Kingdom
Message 1207077 - Posted: 17 Mar 2012, 17:33:29 UTC - in response to Message 1207028.  

As issues go this one won’t cause me to shed many tears, especially so given SIV reported "HKEY_LOCAL_MACHINE\SOFTWARE\Space Sciences Laboratory, U.C. Berkeley\BOINC Setup missing". I did not know BOINC would run if the registry was not setup and given you set all 3 entries to D:\BOINC\ I expect this will only work when both the BOINC programs and client data reside in the same rather than using 2 separate folders. Is this the case?

I am none too sure how I should make SIV work without the registry settings as it does not know where to find BOINC. I guess it could check for BOINC.exe being in the current folder and get the SIV start-up code to create the “missing” registry information. With your setup this would work if you either put all the SIV files in D:\BOINC\ or started SIV with D:\BOINC\ as the “Start in” location.

Would you be happy for SIV to create the registry entries and would the proposed solution be good enough?

If you don’t wish SIV to create the registry keys and values then I expect I could make it work without doing this. It would just be a slightly more pervasive change.

Yeah most issues won't make developers cry unfortunately. :(

BOINC doesn't require and registry entries to exist when you run it. It will make the entry "HKCU\Software\Space Sciences Laboratory, U.C. Berkeley\BOINC Manager", but that just stores the GUI settings for boincmgr.exe I believe. I would have to double check to see if that entry is made on machines where I don't run the manager, which is about 97% of the boxes I use.

When you start BOINC in this manor as there are no settings to tell it where the data directory is located. So it assumes its running directory is the data directory. Probably because that is how it originally worked many moons ago. This is great for how I run BOINC on my machines. Less hassle than having to worry about making registry entries whenever I ghost a machine.

I personally don't like having the entries pointing to the locations as it is an issue when I run other instances of BOINC. If I have say D:\BOINC\ and D:\BOINC2\ on the same machines having entries for either location would cause the other not to function correctly. Also if I need to test something out I will have another instance in another location for whatever I'm doing. While the main one is also running.

I tend to think the best option is look in the registry for the entry and then require input if it can't be found. The rub in that is this isn't dedicated BOINC type software. So you hide the function if not found. Perhaps something like -BOINC_PATH= as a start parameter? As an override to looking in the registry.

OK. How about I just add -PATH= as a start-up qualifier and when specified and it contains BOINC.exe SIV will not look at the registry at all. This will deal with both there being nothing there and also allow for a test version that is in a different folder to the ones specified in the registry. As hope you would expect quotes will be needed for -PATH="C:\Program Files\BOINC\".

Once I make the changes would you like to test a V4.28 Beta? If so then can we sort the details out via e-mail?
ID: 1207077 · Report as offensive
Profile red-ray
Avatar

Send message
Joined: 24 Jun 99
Posts: 308
Credit: 9,029,848
RAC: 0
United Kingdom
Message 1207128 - Posted: 17 Mar 2012, 19:01:56 UTC - in response to Message 1207075.  
Last modified: 17 Mar 2012, 19:27:12 UTC

red-ray wrote:
...
Maybe I should get the SIV [BOINC Status] page to show the current Cricket D/L rate. What do others think?

Scraping data from an HTML page is often frowned upon. Having potentially thousands of hosts running SIV downloading that status just might make it impossible for S@H staff to ever get an agreement with the campus network people allowing more up/download bandwidth by using part of the 1 Gbps link to SSL.

I don't mean to be alarmist, most probably if the question were asked the campus people would say that in the big picture of how much traffic they're handling it would do no harm. But without the question, it could be used as a point against the project.

I suspected there might be concerns, so I will not release a SIV that can do this.

Given the option I would grab XML. I feel it would be good to include the current Cricket information in http://setiathome.berkeley.edu/sah_status.xml.

[BOINC V6.12.34.0 x64 Status] <- SIV64X - System Information Viewer V4.28 Beta-00 RED::ray

        [Y]Cricket 76.45 Mb/sec GPU-0    82°C    81°C    81°C       -  SIV64X is using the -PATH="C:\Program Files\BOINC\" command line qualifier
        [Y]GPU Temperatures     GPU-1    74°C    73°C    72°C    75°C  [_]Launch at Logon   C:\Program Files\BOINC\
        [Y]Auto Retry in 14:48  GPU-2    72°C    71°C    71°C       -  [_]Screen Saver      C:\Program Files\BOINC\
        [Y]CPU Affinity (371)   GPU-3    83°C    82°C    82°C       -  Total WUs 1,869  Other 1,868  Astropulse 1 (0:12)

  PID  CPU Utilisation  Group and CPUs     Name         Version    @    Image Filename - Active 2  Maximum 2  End Limit 895

 4920  Idle              0  CPU-0->CPU-11  |B| boincmgr            8 N  boincmgr.exe
 5764  Idle              0  CPU-0->CPU-11  |B| boinc               8 N  boinc.exe
ID: 1207128 · Report as offensive
Wembley
Volunteer tester
Avatar

Send message
Joined: 16 Sep 09
Posts: 429
Credit: 1,844,293
RAC: 0
United States
Message 1207203 - Posted: 17 Mar 2012, 21:57:35 UTC - in response to Message 1207077.  

OK. How about I just add -PATH= as a start-up qualifier and when specified and it contains BOINC.exe SIV will not look at the registry at all. This will deal with both there being nothing there and also allow for a test version that is in a different folder to the ones specified in the registry. As hope you would expect quotes will be needed for -PATH="C:\Program Files\BOINC\".

Once I make the changes would you like to test a V4.28 Beta? If so then can we sort the details out via e-mail?

Can SIV run with multiple instances with -PATH pointing to different installations of BOINC for each instance?
ID: 1207203 · Report as offensive
Profile red-ray
Avatar

Send message
Joined: 24 Jun 99
Posts: 308
Credit: 9,029,848
RAC: 0
United Kingdom
Message 1207286 - Posted: 18 Mar 2012, 1:06:43 UTC - in response to Message 1207203.  
Last modified: 18 Mar 2012, 1:29:00 UTC

OK. How about I just add -PATH= as a start-up qualifier and when specified and it contains BOINC.exe SIV will not look at the registry at all. This will deal with both there being nothing there and also allow for a test version that is in a different folder to the ones specified in the registry. As hope you would expect quotes will be needed for -PATH="C:\Program Files\BOINC\".

Once I make the changes would you like to test a V4.28 Beta? If so then can we sort the details out via e-mail?

Can SIV run with multiple instances with -PATH pointing to different installations of BOINC for each instance?

Yes that is what I extrapolated as the requirement. SIV V4.28 Beta-00 can/should (aren’t terse responses are great!?), though the only machine in the known world that has this version is http://setiathome.berkeley.edu/show_host_detail.php?hostid=6379672. Would you like to test a Beta? If so then e-mail me directly.
Regards, Ray
ID: 1207286 · Report as offensive
Previous · 1 · 2 · 3

Message boards : Number crunching : SETI@home | Backing off 2 hr 45 min 50 sec on download


 
©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.