KWSN - Chicken of Angnor 的帖子
|
101)
留言板 :
SETI@home Staff Blog :
Eric's biannual post #6: You can tuna fish, but you can't tune a TCP
(消息 571157)
发表于：19 May 2007 作者：  KWSN - Chicken of Angnor
Post: Hi folks, to clear up a common misconception: the optimized apps have nothing at all to do with the current problem. All transfers are handled by BOINC itself, the science apps only crunch what BOINC fetches for them. The current problem lies within the (anonymous) platform mechanism and app_info.xml, the associated settings file. Somehow, the scheduler is erroring out when such a file is present, no matter what application you have in there. Try making an app_info.xml containing the stock app and see for yourself ;o) Also, I've noticed something else that maybe points to a probable cause for our trouble - as at least one other person noticed, I can no longer grab work at all with my 64-bit BOINC client, no matter whether I delete/rename app_info.xml or not. The specific error message looks like this - 2007-05-18 15:03:44 [SETI@home] Sending scheduler request: To report completed tasks which repeats as often as it tries to grab new work and actually gets through to the scheduler. It can't even report its work. So - it really seems like it's a problem with ANY non-standard platform, be it "anonymous" or, like here, "x86_64-pc-linux-gnu". I also tried the official 64-bit BOINC build, but it wants a newer GLIBC than my Gentoo system has to offer, so no go there (it may have a different platform string, haven't checked). Basically, it looks like the recent server code edit has a problem with platforms as a whole, not just with "anonymous" ones. Reverting to 32-bit BOINC (platform string: i686-pc-linux-gnu) got me new work (without app_info.xml, anyway). HTH, Simon. |
|
102)
留言板 :
Number crunching :
Ghost WU issue (and some talk about deadlines)
(消息 571097)
发表于：19 May 2007 作者： KWSN - Chicken of Angnor
Post: Hi folks, to clear up a common misconception: the optimized apps have nothing at all to do with the current problem. All transfers are handled by BOINC itself, the science apps only crunch what BOINC fetches for them. The current problem lies within the anonymous platform mechanism and app_info.xml, the associated settings file. Somehow, the scheduler is erroring out when such a file is present, no matter what application you have in there. Try making an app_info.xml containing the stock app and see for yourself ;o) Also, I've noticed something else that maybe points to a probable cause for our trouble - as at least one other person noticed, I can no longer grab work at all with my 64-bit BOINC client, no matter whether I delete/rename app_info.xml or not. The specific error message looks like this - 2007-05-18 15:03:44 [SETI@home] Sending scheduler request: To report completed tasks which repeats as often as it tries to grab new work and actually gets through to the scheduler. It can't even report its work. So - it really seems like it's a problem with ANY non-standard platform, be it "anonymous" or, like here, "x86_64-pc-linux-gnu". I also tried the official 64-bit BOINC build, but it wants a newer GLIBC than my Gentoo system has to offer, so no go there (it may have a different platform string, haven't checked). Basically, it looks like the recent server code edit has a problem with platforms as a whole, not just with "anonymous" ones. Reverting to 32-bit BOINC (platform string: i686-pc-linux-gnu) got me new work (without app_info.xml, anyway). HTH, Simon. |
|
103)
留言板 :
SETI@home Staff Blog :
Eric's biannual post #6: You can tuna fish, but you can't tune a TCP
(消息 570411)
发表于：18 May 2007 作者： KWSN - Chicken of Angnor
Post: Hi Carl, I also had lots of WUs failing to report before I did this workaround. They did get reported and new ones downloaded afterwards, though. So it did work for me for the same problem, but it may take a bit of luck to get through to the scheduler (took an average of 3 tries for me, but some hosts never got lucky yet). HTH, Simon. |
|
104)
留言板 :
Number crunching :
Revision: 2.3_S5 ??
(消息 570397)
发表于：18 May 2007 作者： KWSN - Chicken of Angnor
Post: Hi Sutaru, that's a pre-release (test) application that is not available to the public yet. It solves the "stuttering" issues people have reported with Rev-2.2B. To illustrate - we use even numbers for release versions (like 2.0/2.2) and odd ones for development ones (2.1/2.3). Pending further testing, a new app release will happen in the next few weeks. HTH, Simon. |
|
105)
留言板 :
Number crunching :
Ghost WU issue (and some talk about deadlines)
(消息 570201)
发表于：18 May 2007 作者： KWSN - Chicken of Angnor
Post: YES!!! I was just about to post the same thing. As illogical as it sounded when I first read it, it seems to be true. Just went through the same procedure on two of my hosts, and voila, results reported, new WUs downloaded, happily crunching again. So I can recommend the above. Now all that remains to be investigated is WHY this happens, and whether we need to adjust all app_info.xml files that are out there. If so, that'd be a major blow because many people will be left without work and without knowing why. Regards, Simon. |
|
106)
留言板 :
Number crunching :
Pay to crunch -- CLOSED
(消息 570173)
发表于：18 May 2007 作者： KWSN - Chicken of Angnor
Post: While I agree that people shouldn't be hammered by requests for donations, the opposite also does not work. Meaning, no one would know you actually need donations to survive, or how to donate even - that is, if noone ever posted about it. So it's more a question of "how much is too much" IMO; people like Crystallize clearly show there is some resentment for whatever reasons - though I do believe such total negativity is misplaced. Regardless, during the current outage donations have been coming in at a much increased rate. The technical situation itself seems to give people enough of a hint. Regards, Simon. |
|
107)
留言板 :
Number crunching :
Pay to crunch -- CLOSED
(消息 569929)
发表于：18 May 2007 作者： KWSN - Chicken of Angnor
Post: In the past 3 days, 11 donations amounting to 310 USD have been made via the PayPal donation gateway at lunatics.at. Not sure about any via the main SETI@Home donation page, but I am sure of the former, since it's my site ;o) I'd also guess it just hasn't been updated recently. Regards, Simon. |
|
108)
留言板 :
Number crunching :
***CLOSED*** SETI/BOINC Milestonesâ„¢ X ***CLOSED***
(消息 569480)
发表于：17 May 2007 作者： KWSN - Chicken of Angnor
Post: Today, the 100th donation to SETI@Home using PayPal was made at lunatics.at. ~2640 USD were donated up til now. Keep up the good work, people! Thank you very much for contributing. Regards, Simon. |
|
109)
留言板 :
SETI@home Staff Blog :
Eric's biannual post #6: You can tuna fish, but you can't tune a TCP
(消息 569019)
发表于：16 May 2007 作者： KWSN - Chicken of Angnor
Post: Well, today some of my hosts managed to upload and report almost all their WUs, vs. an average of 1-2/day/host before. The timeout change certainly seems to have eased the situation somewhat. Still, what Iztok mentioned is worth looking into - unless there is a way for BOINC to recover that WU download, it'll put all low-bandwidth users at a disadvantage while reducing overall project efficiency. Regardless, for a measure in hard times, it's a good one IMO. Regards, Simon. |
|
110)
留言板 :
SETI@home Staff Blog :
Eric's biannual post #6: You can tuna fish, but you can't tune a TCP
(消息 568130)
发表于：15 May 2007 作者： KWSN - Chicken of Angnor
Post:
Yow. Just did that myself three months ago, and lost half the nail. It's grown back since, but damn was that annoying (I type a lot). On the up/download issue, good plan on dropping connections at the router vs. the host itself - hopefully that will have the desired effect and give NFS a kick in the pants. Thanks again for all your and your colleagues' hard work in resurrecting Thumper! Regards, Simon. |
|
111)
疑难解答 :
Web site :
RE: New Private Message Feature.
(消息 568122)
发表于：15 May 2007 作者： KWSN - Chicken of Angnor
Post: Nah, it's new and useful. Go catch up on some sleep, Tony ;o) Regards, Simon. |
|
112)
留言板 :
Number crunching :
UPLOAD/DOWNLOAD Seems working now...
(消息 567843)
发表于：15 May 2007 作者： KWSN - Chicken of Angnor
Post: I have 100's of results I would love to have the kitties send in, but the comms keep failing. The kitties are getting pissed. Won't even look at me, 'cuz they think it's my fault. Have you ever had a kitty give you the evil eye???? Uh oh ;o) Only if I forget to feed them on time....the up/down-load situation is still very patchy. About 10% on average of uploads seem to work for me, no downloads happening for whatever reason. BOINC isn't requesting more work until it can upload its results. So anyway, give Beta/AP/Rosetta etc. a little more CPU time for a little longer maybe? Realistically, I'd say that before Wednesday (after the "usual" outage - not sure whether it'll be skipped) afternoon all bets are off. Regards, Simon. |
|
113)
疑难解答 :
Web site :
RE: New Private Message Feature.
(消息 566641)
发表于：13 May 2007 作者： KWSN - Chicken of Angnor
Post: Yup, I read that one. The outbox wasn't mentioned there though, and I think the abuse potential is pretty high, so it should be simple to set up restrictions. Regards, Simon. |
|
114)
疑难解答 :
Web site :
RE: New Private Message Feature.
(消息 566630)
发表于：13 May 2007 作者： KWSN - Chicken of Angnor
Post: ...maybe even an outbox? :o) It should also have a link to be able to ignore users quickly; not all such messages may be wanted, and moderation is almost impossible because it would take too much effort. Regards, Simon. |
|
115)
留言板 :
Number crunching :
A enterprise class server system for SETI@home?
(消息 566444)
发表于：13 May 2007 作者： KWSN - Chicken of Angnor
Post: That's what I meant "durch die Blume" ;o) Being a user-driven thing, I'd think that would self-regulate - though I agree careful selection should take place. Regards, Simon. |
|
116)
留言板 :
Number crunching :
A enterprise class server system for SETI@home?
(消息 566186)
发表于：13 May 2007 作者： KWSN - Chicken of Angnor
Post: In that case, all negotiations would need to be handled by someone who is not a staff member. Still, I'm sure there are people who could (and maybe even would) do this. It would really help to make the "wish list" in regards to hardware that much more specific. Here's a wild idea: we could even make a "friends of SETI@Home" not-for-profit organization to be able to handle such things now and for the future - meaning readily available contact, and no huge operating restrictions. Maybe that's aiming too high, but I could see that working out to all our advantage. Regards, Simon. |
|
117)
留言板 :
Number crunching :
A enterprise class server system for SETI@home?
(消息 566149)
发表于：12 May 2007 作者： KWSN - Chicken of Angnor
Post: Thing is, you can do the same thing Sun did with that server and use multiple straight SATA cards - these come in all sorts of form factors and PCI/PCI-X/PCI-Express variants with up to 6 ports at low cost (many more for more cash). So plugging 4 of these into any system capable of hosting that many cards should work. In a pinch, you could even use port replicators - these make two or four SATA ports out of one. Reduced bandwidth, but it works. Cheap as chips, as they say. Still, this is all idle musing. No matter whether the project staff themselves are allowed to approach anyone or any company directly - I'm sure there's someone who could maybe, say, handle the introductions? :o) Regards, Simon. |
|
118)
留言板 :
Number crunching :
A enterprise class server system for SETI@home?
(消息 566127)
发表于：12 May 2007 作者： KWSN - Chicken of Angnor
Post: Ned, actually, yes. They did say they used generic Linux software raid, so as long as the drives are recognized by the OS, you can re-assemble the filesystem on any Linux-capable hardware. So anyway, it obviously made for a very speedy recovery having the same chassis, but it was not a necessity. Still, why not ask Dell now? We could still very much use a backup for the Master Science DB ;o) Regards, Simon. |
|
119)
留言板 :
Number crunching :
Change on Server Status?
(消息 566116)
发表于：12 May 2007 作者： KWSN - Chicken of Angnor
Post: A couple of my hosts already got lucky and resumed crunching as of a few minutes ago. Great work, S@H team :o) Feed those kitties, now... Happy crunching, Simon. |
|
120)
留言板 :
Technical News :
Takin' Care of Business and Workin' Overtime (May 11 2007)
(消息 566048)
发表于：12 May 2007 作者： KWSN - Chicken of Angnor
Post: Thank you Sun Microsystems, and thank you very much SETI@Home staff for the very quick recovery with the newly arrived Thumper chassis. Great job in a difficult situation! Regards, Simon. |
前 20 · 后面 20
©2020 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.