Message boards :
Number crunching :
Test post..
Message board moderation
Author | Message |
---|---|
Matt Lebofsky Send message Joined: 1 Mar 99 Posts: 1444 Credit: 957,058 RAC: 0 |
This is a test post. Try your best to ignore it. - Matt - SETI@home/BOINC BOINC/SETI@home |
Petit Soleil Send message Joined: 17 Feb 03 Posts: 1497 Credit: 70,934 RAC: 0 |
> This is a test post. Try your best to ignore it. > > - Matt - SETI@home/BOINC > BOINC/SETI@home > I can't ignore it ! I have a crucial question. Since the data has been backed up what should we do with the WU we have downloaded last time ? Are these WU still "good to process" Thanks in advance. I am sure I am not the only one wondering. Regards Marc -.-. --.- -.. -..- . - .-.-. -.- --... ...-- |
Matt Lebofsky Send message Joined: 1 Mar 99 Posts: 1444 Credit: 957,058 RAC: 0 |
This should be in its own message board thread, but I'll post some info about this later. Basically, since the database is "old" it is completely out of sync with all the workunits that have been downloaded since August 11th. All those clients will eventually process these workunits, try to return the result, and then fail. Instead of wasting this time, we will ask (once the project is back up) that all users restart the project (you won't lose old credit for workunits processed before August 11th, only for ones processed since then). Yes this is bad. I know. But this is the best we can do when we encounter a hardware problem that destroys the database and we have to recover from a backup copy. - Matt - SETI@home/BOINC BOINC/SETI@home |
Petit Soleil Send message Joined: 17 Feb 03 Posts: 1497 Credit: 70,934 RAC: 0 |
> Yes this is bad. I know. But this is the best we can do when we encounter > a hardware problem that destroys the database and we have to recover from > a backup copy. Thanks Matt at least we know and we won't waste our time crunching our current queue. regards Marc -.-. --.- -.. -..- . - .-.-. -.- --... ...-- |
Bruno Moretti IK2WQA Send message Joined: 15 May 99 Posts: 284 Credit: 49,167 RAC: 0 |
Matt Lebofsky wrote: > Yes this is bad. I know. But this is the best we can do when we encounter > a hardware problem that destroys the database and we have to recover from > a backup copy. > > - Matt - SETI@home/BOINC > BOINC/SETI@home Thanks you very much for info, Matt! [/url] |
Byron Leigh Hatch @ team Carl Sagan Send message Joined: 5 Jul 99 Posts: 4548 Credit: 35,667,570 RAC: 4 |
|
Arm Send message Joined: 12 Sep 03 Posts: 308 Credit: 15,584,777 RAC: 0 |
Hi, Matt I have downloaded few WUs before 11th but part of them I have uploaded AFTER 11th. What then? 10x |
eL_nino Send message Joined: 11 Mar 04 Posts: 79 Credit: 999,964 RAC: 0 |
> Hi, Matt > I have downloaded few WUs before 11th but part of them I have uploaded AFTER > 11th. What then? > 10x > What about WU that were downloaded before August 11th and returned them on around August 15th?!?! PLease answer... |
Scottatron Send message Joined: 15 Jul 03 Posts: 94 Credit: 220,389 RAC: 0 |
My God! What sort of rubbish is this? Tell me how Berkeley cares about their users when they don't even dedicate a post on the main page! I bet they would've known about it too, but instead of facing humiliation (When posting on the main board), they just keep it nice and quiet until someone asks! Frankly this is extremely disappointing. I think the below sums it all up nicely, taking from OcUK's own MikeSW17: I can hardly believe what I'm reading What kind of RAID are Berkeley using? A pretty poor one if it allowed a hardware/firmware failure to corrupt data. Perhaps Snap Appliance would like to comment in defence of their product? What kind of SQL database are Berkeley using? A pretty poor one if it caused data corruption that it didn't detect and roll-back the offending transactions. It's very hard not to become totally disalusioned and depressed at the way BOINC's going. Not to mention questioning the quality and abilities of the design/development team. ############################################################################### At least Predictor are open and honest with their users! |
EclipseHA Send message Joined: 28 Jul 99 Posts: 1018 Credit: 530,719 RAC: 0 |
> My God! > > What sort of rubbish is this? Tell me how Berkeley cares about their users > when they don't even dedicate a post on the main page! > At least Predictor are open and honest with their users! > Does Seti/C2 ever admit that they caused a problem? Seem's something else (HW or whatever) made dev "jump"... Predictor, does a very respectable job of letting the "super computer" know what's happening, and Seti/C2 should take some lessons.... Case in point was the "Matt" post. Matt's one of the main-line Boinc/Seti people, and while his message was in a few threads in this forum, he then backed up and stated "It's not offical".. Heck Matt is the webmaster of this site! Who else can post an "official" notice! All I know, is that one or more messages should be sent to the marketing folks at SNAP, asking why their "top of the line" box at UCB crapped out so badly that UCB had to use an 8 day old tape - after one day!..... (there URL is available on the Classic site.. Under sponsors!) |
Matt Lebofsky Send message Joined: 1 Mar 99 Posts: 1444 Credit: 957,058 RAC: 0 |
I'm webmaster in that I start/stop apache. Everybody on the (very small) team contributes to the content. I won't post anything on the front page unless everybody (especially the project directors) agrees with it. My comments here informal notices to those seeking a quick "inside scoop" but I may be completely off the mark for all I know. Hence "unofficial." Anyway, we're actively working on improving the lack-of-PR situation. And it actually wasn't an 8-day-old tape. The database was last updated 8 days ago (because the project was down). So in reality the data is only 3-4 days old. Well, except for the message board data. And it wasn't even tape. It was a copy of the database we still had on line the whole time. We decided since we were down to reformat the raid device the old database was on for better throughput. I think we gained 2.5x the write speed in the process. That remains to be seen. - Matt BOINC/SETI@home |
Byron Leigh Hatch @ team Carl Sagan Send message Joined: 5 Jul 99 Posts: 4548 Credit: 35,667,570 RAC: 4 |
|
StokeyBob Send message Joined: 31 Aug 03 Posts: 848 Credit: 2,218,691 RAC: 0 |
Matt, Thanks for the messages you post. Official or unofficial they are always informative. Keep up the good work. Carson |
slavko.sk Send message Joined: 27 Jun 00 Posts: 346 Credit: 417,028 RAC: 0 |
Thank you for "unofficial", anyway. Keep posting ;o))) > I'm webmaster in that I start/stop apache. Everybody on the (very small) team > contributes to the content. I won't post anything on the front page unless > everybody (especially the project directors) agrees with it. My comments here > informal notices to those seeking a quick "inside scoop" but I may be > completely off the mark for all I know. Hence "unofficial." Try WAP user info http://slavko.sk/wap/seti |
EclipseHA Send message Joined: 28 Jul 99 Posts: 1018 Credit: 530,719 RAC: 0 |
> I'm webmaster in that I start/stop apache. Everybody on the (very small) team > contributes to the content. I won't post anything on the front page unless > everybody (especially the project directors) agrees with it. So, you send -HUP to main apache process? If it's a "very small" team, seems it should be easy to get thier buy in... Now the "project directors", do they have the technical knowledge of the problems, to make a call? Ya, they might change the "wording", but on technical content, when you post in a forum, it's the "word of UCB" > Anyway, we're actively working on improving the lack-of-PR situation. Step 1 - get the "project directors" out of the loop, as they have no clue on the technology that the "supercomputer" is dealing with. They might noteven understand the technology you're dealing with on the server end. Step 2 - Look at predictor as an example. The Dev team has been telling the "supercomputer" when they messed up, and all in all, has been taken well! Be able to admit your problems! > > And it actually wasn't an 8-day-old tape. The database was last updated 8 days > ago (because the project was down). So in reality the data is only 3-4 days > old. Seems to me, that STEP1 of moving the DB from the old DB to the new one (as the project was down) would be to BACKUP the DB! Be it on tape, or another disk > Well, except for the message board data. And it wasn't even tape. It was a > copy of the database we still had on line the whole time. We decided since we > were down to reformat the raid device the old database was on for better > throughput. I think we gained 2.5x the write speed in the process. That > remains to be seen. Let's see.. disk-to-disk or tape-to-disk copy... I think I know here I'd put my money on "which was faster"! (only 2.5?) |
bjacke Send message Joined: 14 Apr 02 Posts: 346 Credit: 13,761 RAC: 0 |
|
The Gas Giant Send message Joined: 22 Nov 01 Posts: 1904 Credit: 2,646,654 RAC: 0 |
But the post isn't yellow...... |
bjacke Send message Joined: 14 Apr 02 Posts: 346 Credit: 13,761 RAC: 0 |
|
Mjöllnir Send message Joined: 3 Apr 99 Posts: 46 Credit: 0 RAC: 0 |
AZ Woody AKA conceited Pea Cock God "I am Napoleon/Einstein of SETI@home, I am better than all Berkeley, Anderson newbie, vote AZ Woody for SETI@home direction!!!" |
Toby Send message Joined: 26 Oct 00 Posts: 1005 Credit: 6,366,949 RAC: 0 |
AZ-Woody: Well that proves it. You have NO CLUE what you are talking about. We have all suspected it before but right there is the proof I have been waiting on. > Step 1 - get the "project directors" out of the loop, as they have no clue on > the technology that the "supercomputer" is dealing with. They might noteven > understand the technology you're dealing with on the server end. You might want to do a google search before insulting people. I quote from the seti-classic page: "Project Director. David [Anderson] has done extensive research in operating systems and distributed computing He designed the SETI@home software and data architecture." He just designed the infrastructure for the world's largest super computer. He knows nothing. The director is one of the devs! He has contributed much code to BOINC. Also, may I point out that you are being a hypocritical, stuck up snob? First you want more info and then you complain about it when it comes. I actually agree with you on the communications issue. It would be nice to hear more detailed news more often. But at least I say "thanks" when we get some. Matt (and Rom) is trying to calm some nerves with his 'unofficial' remarks. He takes time out of his day to post here for our benefit. A lot of times the rest of the team may be busy or unavailable so he can't get a consensus and make it 'official' but he still tells us what he knows. Were you raised by wolves or have you just forgotten how to be polite? Try it: "Thank you sir, may I have another?" There... was that so hard? ------------------------------------------- - A member of The Knights Who Say NI! Possibly the best stats site in the universe: http://boinc-kwsn.no-ip.info |
©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.