Hi. Boinc worked fine until last job was done. Now it says "needs more space".

IMG: http://www.zeitgeist3.dk/images/boinc_prob_01.jpg

Above 100gb free on all drives, no other app running (besides antivirus + peerblock, which worked fine so far).

OS: XP Pro

MEM: 6Gb (4gb used under XP)

Boinc ver: 6.10.58
Disk usage set for:
30
0,10
50,00
60
75,00

Hi. Boinc worked fine until last job was done. Now it says "needs more space".

IMG:

I displayed the image. You could have copy&pasted the lines here (albeit without colour;-).

What does BOINC report in the Messages tab after a restart (reboot) for "Preferences limit disk usage to"?

What values does BOINC show in the Disk tab?

Gruß,
Gundolf

---

Computer sind nicht alles im Leben. (Kleiner Scherz)

>What does BOINC report in the Messages tab after a restart (reboot) for "Preferences limit disk usage to"?
22-02-2011 10:12:04 Starting BOINC client version 6.10.58 for windows_intelx86
22-02-2011 10:12:04 log flags: file_xfer, sched_ops, task
22-02-2011 10:12:04 Libraries: libcurl/7.19.7 OpenSSL/0.9.8l zlib/1.2.3
22-02-2011 10:12:04 Data directory: D:\Documents and Settings\All Users\Application Data\BOINC
22-02-2011 10:12:04 Running under account jesper
22-02-2011 10:12:04 Processor: 8 GenuineIntel Intel(R) Core(TM) i7 CPU 920 @ 2.67GHz [Family 6 Model 26 Stepping 4]
22-02-2011 10:12:04 Processor: 256.00 KB cache
22-02-2011 10:12:04 Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss htt tm pni ssse3 cx16 sse4_1 sse4_2 nx lm vmx tm2 popcnt pbe
22-02-2011 10:12:04 OS: Microsoft Windows XP: Professional x86 Edition, Service Pack 3, (05.01.2600.00)
22-02-2011 10:12:04 Memory: 3.00 GB physical, 4.84 GB virtual
22-02-2011 10:12:04 Disk: 488.28 GB total, 303.44 GB free
22-02-2011 10:12:04 Local time is UTC +1 hours
22-02-2011 10:12:04 NVIDIA GPU 0: GeForce GTX 260 (driver version 26099, CUDA version 3020, compute capability 1.3, 896MB, 537 GFLOPS peak)
22-02-2011 10:12:04 AQUA@home URL http://aqua.dwavesys.com/; Computer ID 94149; resource share 100
22-02-2011 10:12:04 No general preferences found - using BOINC defaults
22-02-2011 10:12:04 Reading preferences override file
22-02-2011 10:12:04 Preferences:
22-02-2011 10:12:04 max memory usage when active: 1535.12MB
22-02-2011 10:12:04 max memory usage when idle: 2763.21MB
22-02-2011 10:12:04 max disk usage: 30.00GB
22-02-2011 10:12:04 don't use GPU while active
22-02-2011 10:12:04 suspend work if non-BOINC CPU load exceeds 25 %
22-02-2011 10:12:04 (to change preferences, visit the web site of an attached project, or select Preferences in the Manager)
22-02-2011 10:12:04 Not using a proxy
22-02-2011 10:12:04 AQUA@home Sending scheduler request: To fetch work.
22-02-2011 10:12:04 AQUA@home Requesting new tasks for GPU
22-02-2011 10:12:06 AQUA@home Scheduler request completed: got 0 new tasks
22-02-2011 10:12:06 AQUA@home Message from server: No work sent
22-02-2011 10:12:17 AQUA@home Sending scheduler request: To fetch work.
22-02-2011 10:12:17 AQUA@home Requesting new tasks for CPU
22-02-2011 10:12:19 AQUA@home Scheduler request completed: got 0 new tasks
22-02-2011 10:12:19 AQUA@home Message from server: No work sent
22-02-2011 10:12:19 AQUA@home Message from server: D-Wave's Redfield Open Quantum Simulation: Multi-Threaded needs 275.83MB more disk space. You currently have 2108.36 MB available and it needs 2384.19 MB.
22-02-2011 10:21:45 AQUA@home Sending scheduler request: To fetch work.
22-02-2011 10:21:45 AQUA@home Requesting new tasks for CPU
22-02-2011 10:21:47 AQUA@home Scheduler request completed: got 0 new tasks
22-02-2011 10:21:47 AQUA@home Message from server: No work sent
22-02-2011 10:21:47 AQUA@home Message from server: D-Wave's Redfield Open Quantum Simulation: Multi-Threaded needs 275.86MB more disk space. You currently have 2108.33 MB available and it needs 2384.19 MB.
22-02-2011 10:38:05 AQUA@home Sending scheduler request: To fetch work.
22-02-2011 10:38:05 AQUA@home Requesting new tasks for GPU
22-02-2011 10:38:07 AQUA@home Scheduler request completed: got 0 new tasks
22-02-2011 10:38:07 AQUA@home Message from server: No work sent

What values does BOINC show in the Disk tab?
http://www.zeitgeist3.dk/images/boinc_prob_02.jpg

This is a problem which appeared at AQUA immediately after they upgraded their server to correct some user web page errors.

In updating, they also happened to catch changeset [trac]changeset:23079[/trac], which I suspect is implicated in this problem. Last I heard (http://aqua.dwavesys.com/forum_thread.php?id=747&nowrap=true#10502), they were trying to revert to an earlier version, but they don't appear to have succeeded.

Also, last night I couldn't get any mail through to Berkeley (my ISP didn't have a DNS MX record for them), so the developers may not be aware yet. It may be several more hours before anybody starts working on this problem.

In updating, they also happened to catch changeset [trac]changeset:23079[/trac], which I suspect is implicated in this problem.

Changeset [trac]changeset:23079[/trac] was a response to my complaint that the client is mixing quota-limited part of disk size, available to user, with total partition free space. As it is a Windows code (intended for the client), I hope it should have no influence on the server...

Last I heard (http://aqua.dwavesys.com/forum_thread.php?id=747&nowrap=true#10502), they were trying to revert to an earlier version, but they don't appear to have succeeded.

...unless the client would report some ill values back to server, but AFAIK it was applied just on trunk past the 6.12.15 tags.

Peter

In updating, they also happened to catch changeset [trac]changeset:23079[/trac], which I suspect is implicated in this problem.

Changeset [trac]changeset:23079[/trac] was a response to my complaint that the client is mixing quota-limited part of disk size, available to user, with total partition free space. As it is a Windows code (intended for the client), I hope it should have no influence on the server...

Last I heard (http://aqua.dwavesys.com/forum_thread.php?id=747&nowrap=true#10502), they were trying to revert to an earlier version, but they don't appear to have succeeded.

...unless the client would report some ill values back to server, but AFAIK it was applied just on trunk past the 6.12.15 tags.

Peter

The reported errors are all of the "message from server" type. I think it's highly unlikely that so many people would both compile their own v6.12.16 clients, and yet not mention it when posting.

It seems to affect exclusively ROQS tasks, which are huge - in both memory and time. I should have one finishing in a couple of hours (after 24 CPU-days!), so I'll try and fetch a replacement (AQUA limits to two tasks per host, and I have my spare already). That should give me both some logs, and a matched pair of sched_request and sched_reply files to pore over.

In updating, they also happened to catch changeset [trac]changeset:23079[/trac], which I suspect is implicated in this problem. Last I heard (http://aqua.dwavesys.com/forum_thread.php?id=747&nowrap=true#10502), they were trying to revert to an earlier version, but they don't appear to have succeeded.

...unless the client would report some ill values back to server, but AFAIK it was applied just on trunk past the 6.12.15 tags.

The reported errors are all of the "message from server" type. I think it's highly unlikely that so many people would both compile their own v6.12.16 clients, and yet not mention it when posting.

Indeed...

It seems to affect exclusively ROQS tasks...

...what all makes me think that the AQUA server is solely responsible for this.
Just which changeset since their update did it... (just definitely no guinea-pigs testing changeset [trac]changeset:23079[/trac] ;-).

Peter

It seems to affect exclusively ROQS tasks...

...what all makes me think that the AQUA server is solely responsible for this.
Just which changeset since their update did it... (just definitely no guinea-pigs testing changeset [trac]changeset:23079[/trac] ;-).

Peter

Not many to choose from. They were at least at [23007] beforehand, because that one was written to fix a bug which AQUA themselves encountered and reported - and the fix worked for two weeks, until last night's upgrade.

I've got those logs and request/reply now - trying to make sense of them before I try posting to boinc_alpha.

I went through all the entries from 23007 to 23079, unless one of the preferences changes did it, there hasn't been anything else. Yet I agree with Peter that 23079 is for the client only and so you will see this change in 6.12.16. It shouldn't affect any server.

AQUA forums are now gone/off line? All I see is white. ;)

---

I went through all the entries from 23007 to 23079, unless one of the preferences changes did it, there hasn't been anything else. Yet I agree with Peter that 23079 is for the client only and so you will see this change in 6.12.16. It shouldn't affect any server.

AQUA forums are now gone/off line? All I see is white. ;)

Front page is live, but if I try to navigate from there to the message boards I get an HTTP 500 error. Sounds like the server is well scrambled.

Looks as if AQUA is trying to send multiple WUs, sometimes even multiple tasks from the same WU. My sched_reply file has 14 </workunit> tags.

Since the ROQS tasks are huge, and the limit should be 2 per host (and I had one already, so should have been limited to 1 this time), that might explain it.

The sched_reply also contains

<message priority="notice">D-Wave's Redfield Open Quantum Simulation: Multi-Threaded needs 1414.81MB more disk space. You currently have 969.37 MB available and it needs 2384.19 MB.</message>

I'll try and email them through, but my ISP can't find Berkeley at the moment.

Forwarded to development.

---

AQUA's scheduler has resumed sending out the smaller Fokker-Planck tasks - they also failed last night, with a message about the application not being permitted by preferences ("No work available for the applications you have chosen").

Edit - and now the message boards are up as well.

Seems to be working now. First packet after it picked up again, was very small (disk usage: 66mb) compared to last packet (disk usage 1002mb) processed.

Thank you all, for looking into this problem, being a part of fixing it or not :)

Log
22/02/2011 20:34:37 AQUA@home Requesting new tasks for CPU
22/02/2011 20:34:39 AQUA@home Scheduler request completed: got 1 new tasks
22/02/2011 20:34:41 AQUA@home Started download of instance.txt_roqs_16feb11_a_p1_030_25000_1_4
22/02/2011 20:34:41 AQUA@home Started download of instance.txt_roqs_16feb11_a_p1_030_25000_1_4.noise
22/02/2011 20:34:42 AQUA@home Finished download of instance.txt_roqs_16feb11_a_p1_030_25000_1_4
22/02/2011 20:34:42 AQUA@home Started download of instance.txt_roqs_16feb11_a_p1_030_25000_1_4.hamiltonian
22/02/2011 20:34:43 AQUA@home Finished download of instance.txt_roqs_16feb11_a_p1_030_25000_1_4.noise
22/02/2011 20:34:45 AQUA@home Finished download of instance.txt_roqs_16feb11_a_p1_030_25000_1_4.hamiltonian
22/02/2011 20:34:45 AQUA@home Starting roqs_16feb11_a_p1_030_25000_1_4_0
22/02/2011 20:34:45 AQUA@home Starting task roqs_16feb11_a_p1_030_25000_1_4_0 using ROQS version 14

Yeah, but now they're on [trac]changeset:23080[/trac]. I'd really like Richard's explanation on how a compiler fix for the client fixed the problem at AQUA. ;-)

---