Problems in the server
Message boards :
News :
Problems in the server
Message board moderation
Author | Message |
---|---|
Send message Joined: 18 Mar 15 Posts: 282 Credit: 2,748,608 RAC: 0 |
Dear volunteers, These days we have been suffering problems in the database due to an update. Many information was corrupted in the process. We have recovery a backup copy but it is a bit old. We are working to get everything working well as soon as possible. Please, be patient and thank you in advance. Best regards. Jesús Carro Universidad San Jorge @InSilicoHeart |
Send message Joined: 7 Nov 15 Posts: 11 Credit: 29,828,122 RAC: 1 |
My stats have dropped down to 2016 level..... did have 3.186 million, now showing 1.925 million :( |
Send message Joined: 18 Aug 15 Posts: 3 Credit: 122,200 RAC: 0 |
I had running one task and another few waiting for run, I clicked "Update" in BOINC manager and all tasks from DENIS has been cancelled. |
Send message Joined: 9 Apr 15 Posts: 14 Credit: 328,091 RAC: 0 |
You have deleted months of work. Michael. President of Rechenkraft.net, Principal Investigator of the RNA World distributed supercomputer. |
Send message Joined: 8 Apr 15 Posts: 34 Credit: 389,238 RAC: 0 |
I HAD 94,882 credits the day before the crash and lost at least 66,781 according to my BOINCSTATS TRACKING PAGE.[url][/url] |
Send message Joined: 13 Apr 15 Posts: 13 Credit: 1,590,484 RAC: 0 |
There are people who have lost more than 1,000,000 credits, look it up in BOINCStats and/or FreeDC... |
Send message Joined: 13 Apr 15 Posts: 13 Credit: 1,590,484 RAC: 0 |
Dear volunteers, Next time you update, do a backup first... In the meantime: Good luck with the restore, I hope the project hasn't lost data. |
Send message Joined: 9 Apr 15 Posts: 5 Credit: 651,184 RAC: 0 |
hear hear !!! lesson no. 1: make a backup BEFORE any routine OR important updates lesson no. 2: make sure the backup can be restored (onto a spare machine if needs be). lesson no.3: rotate backups on a very regular basis - if your last backup is 6+months old, your disaster recovery procedures need rewriting and your IT dept head should be changed. I hope people will be understanding...but losing many peoples time and money (electricity bills, hardware etc) makes recommending this project more difficult if the data is lost due to negligence. Such a shame :-( Hope you can resolve the issues. Tim Founder, UK BOINC Team |
Send message Joined: 9 Apr 15 Posts: 14 Credit: 328,091 RAC: 0 |
You need a DAILY backup of your entire data and you need to rotate storage of these backups such that potentially bad data doesn't overwrite a good backup. You did not even check your server status for at east 2 days because otherwise you would have noticed much earlier that your entire website was gone, too. Obviously, you have not made backups for more than 6 months as can be derived from the stats loss. All OUR energy ressources which we invested in your research progress have therefore been wasted if you do not come up with a backup. The least one can expect, however, is a clean restoration of the credits plus badges which all have gone. And one more thing: I expect you to STANDARDLY implement the optimized apps for all platforms to stop wasting our energy ressources! The changes to your original code and its testing is well documented and consists of the implementation of profoundly more professional code compared to your original (e.g. GROMACS floating point operations routines, if I remember correctly). For many years I have witnessed this igorance of, what I like to call, "ivory tower researchers" working in multiple research fields who on the one hand believe, they have all the wisdom in the world while - in fact - they are not even capable of evaluating a code optimization which is suggested to them free of charge and would boost their research throughput profoundly while saving our energy. I met high energy physicists who laugh at the distributed computing community and continue to waste tax payers money for high performance cluster setup and maintenance not realizing that these HPC, while suitable (and sometimes even required) for some (limited types of) tasks, are basically a "fly's fart in the wind" compared the compute power the DC community offers for free. The same is true for many biotechnologists. If you don't comply to the above mentioned requirements, you will probably need a vast amount of money in the very near future to re-fund the re-gathering of all the data that now was lost - then using a costly HPC system. Because we will be out. Michael. President of Rechenkraft.net, Principal Investigator of the RNA World distributed supercomputer. |
Send message Joined: 8 Apr 15 Posts: 34 Credit: 389,238 RAC: 0 |
Until things have finally been FULLY RESTORED to where they were pre-database/server crash I have decided to QUIT SUPPORTING THIS PROJECT with my time and electricity. As said by UBT - Timbo in Message 1242 "I hope people will be understanding...but losing many peoples time and money (electricity bills, hardware etc) makes recommending this project more difficult if the data is lost due to negligence." |
Send message Joined: 22 May 15 Posts: 2 Credit: 8,633,199 RAC: 0 |
2 days ago, I had approximately 8,190,000 credits. Now I have only 3,659,616 ... Right now my computers have about 200 work units pending upload or report, but they seem to be not on the Task List Web Page I'm doubt I can continue to contribute to this project or not ... |
Send message Joined: 2 Oct 16 Posts: 7 Credit: 1,038,237 RAC: 0 |
I has about 900k and all this has been wiped out completely :( I hope you will be able to fix this. |
Send message Joined: 18 Mar 15 Posts: 282 Credit: 2,748,608 RAC: 0 |
Dear all, I give you some more information to try to clarify some issues regarding this problem. All the results of the simulations are safe. Every hour they are uploaded to Google Drive. Only the workunits of the last three days could be lost. Your statistics are stored each day in an xml file. We'll be able to solve the problem with your credit but we need time to develop a script to do it. We have an script to do backup every day and to store it in two different servers. I don't know yet why it has not been working. I have to find out. In parallel, I'm talking with the IT department to improve the backup system. I fully understand everything you are saying and how you feel. Please also understand my surprise when I discovered that things that would have to be doing well were not working. I ask again patience. Give it for sure that I will do everything in my hand to try to fix it. Best regards, Jesús Jesús Carro Universidad San Jorge @InSilicoHeart |
Send message Joined: 9 Apr 15 Posts: 14 Credit: 328,091 RAC: 0 |
Thank you very much for this more detailed feedback. After this, I think, the situation might be much better than I had initially thought. I hope you will be able to solve the problems fast. But I stick to my request to deliver the improved apps as standard apps directly from the DENIS@home website: A speed increase to you means an increase in results throughput. To us it means a reduction in electricity costs and hardware stressing. And as far as I remember, that speed increase was more than just significant. Michael. President of Rechenkraft.net, Principal Investigator of the RNA World distributed supercomputer. |
Send message Joined: 22 May 15 Posts: 2 Credit: 8,633,199 RAC: 0 |
Thank you for the information. As the simulation result files are safe, I feel relieved. Good luck for the recovery, and hope we can continue the science! |
Send message Joined: 2 Oct 16 Posts: 7 Credit: 1,038,237 RAC: 0 |
Thanks for update! BTW, if you could not recover stats from xml files, as a last resort you could try to load data from external stats sites. |
Send message Joined: 9 Apr 15 Posts: 5 Credit: 651,184 RAC: 0 |
Dear all, Hi Jesus Excellent news - if only a small fraction of the data has been lost then this makes for a better situation. 3 days loss is much much better than 6+ months !!
For me, I think the stats are less important - sure they are nice to know how much each member contributes, but it is the science that is important. If you can fix some of the "knock on" effects from the server crash then all the better. This will at least restore some peoples faith in the project. I would even suggest that you suspend the project for a week, so you have time to get any fixes in place and apply them. regards and good luck with all of this ! Tim |
Send message Joined: 10 Oct 16 Posts: 18 Credit: 14,819,265 RAC: 0 |
Real crunchers are not here to get credits, so just stop crying about something such unimportant as that. Real crunchers are here to support science and in the case of DENIS@home - at least in my opinion - we support the most important section of science: medical research. That the project has maybe lost some data and was down for a few days, well such things just happen. As soon as the servers are on again, we will push this project further than it ever was before :-) |
Send message Joined: 7 Nov 15 Posts: 11 Credit: 29,828,122 RAC: 1 |
Real crunchers are not here to get credits, so just stop crying about something such unimportant as that. Real crunchers are here to support science and in the case of DENIS@home - at least in my opinion - we support the most important section of science: medical research. That the project has maybe lost some data and was down for a few days, well such things just happen. As soon as the servers are on again, we will push this project further than it ever was before :-) Well said!!!! My contribution will continue..... the science matters, the BOINC credit does not!!! |
Send message Joined: 7 Jul 15 Posts: 28 Credit: 31,154,473 RAC: 0 |
To Thomas as well as TheFiend... Ever notice that items seem to bring a better price at auctions rather than a simple flea market sale? More participation and contests promote more crunching!!! BOINC Points are a measure of contribution relative to other contributors. Having points and teams and competitions ONLY HELP the cause! I think it's rather narrow minded and unfair to say "the science matters, the BOINC credit does not!!!" when it does in fact promote MORE participation. "Real crunchers are not here to get credits, so just stop crying about something such unimportant as that." WOW and WOW! So I am not a real cruncher? Others are not real crunchers? Judge Jury Executioner you are? I have 13 setups running at the moment and I care about points and I am not a real cruncher? Good grief I hope nobody believes that statement!! Y'all suppose anybody who cares about points should just shut up and BOINC or not without any comments? Doesn't that viewpoint promote alienation and possibly cause folks to participate less? Points matter; maybe not as much as the science involved, but it's far from zero and is a valid complaint subject. 8-) |