March 5, 2009
The time you have to return results for new workunits has been
decreased to 24 hours. The reason is that we try to minimize
the number of workunits being in progress as we will upgrade
the server very soon.
February 6, 2009
The project will be down for a few hours on wednesday, Feb 18th because some
maintenance tasks on our USV are planned.
November 11, 2008
There were 1261 workunits in a permanent credit-pending state. The
problem should be solved now. Don't hesitate to use the fora if it
happens again.
October 21, 2008
Sorry for the unexpected downtime. Our SCSI disks are quite small and
they simply went out of space. [x] fixed. Currently the server fights
with a huge backlog but in a few hours the situation should normalize
again.
August 25, 2008
Database trouble. Please cancel all workunits and reset your clients
immediately. See the news section for the reason.
August 24, 2008
This weekend our database became very slow and users weren't able to
contact the server most of the time. Sorry. The problem should be
fixed soon. If you have any rcross* workunits left, please cancel
them immediately. Thx!
August 14, 2008
We are inserting a new bunch of 1 167 876 W7 workunits right now.
They should keep your machines busy for the next 3 months :-]
August 11, 2008
We have some trouble with the tcape application and right now I'm
trying to solve the problem. The server might become unavailable for
some time today. Sorry.
July 15, 2008
Tomorrow there will be a power cut-off at our institute. Therefore all
our servers will be shut down from today in the evening until wednesday
noontime. Moreover I will be out of office for the rest of the week.
However, I will keep an eye on the Boinc servers. Best regards, Bernhard.
July 14, 2008
Sorry for the unexpected downtime. We had some trouble with our network
this weekend. Now it works again, but note that we will need to shut down
our servers for a few hours on wednesday.
July 7, 2008
Today we had trouble with bad workunits. The problem is solved now. Sorry
for the downtime.
May 21, 2008
Lust for W7? Very soon I will insert the first W7s. I will lower the
number of workunits each host can get _temporarily_. This will speed
up the first results such that we can see if everything works as
intended. If there are any problems, don't hesitate to use the
fora.
April 3, 2008
20 workunits named ''test_*'' are online now. The new series will follow
soon.
March 31, 2008
There are no W6 anymore. We've canceled them in order to do an interim
analysis. The remaining work will be splitted into smaller parts to
increase the efficiency of the project. You can choose to crunch down
(within 2 days) or cancel the W6's you have. For the next few days
we will crunch rCross. Then the W7 series will start.
January 24, 2008
Good news: RCN finally has a screensaver. Johannes told me yesterday that
it works for Windows and 32-bit Linux. The screensaver has been implemented
by our user ''S@NL FilmFreak''. Thank you, Robin!
January 9, 2008
Please cancel all the old workunits, currently we crunch the W0q series for
testing the system. I will inform you asap when the error is eliminated.
January 9, 2008
We have some trouble with the new workunits and currently we debug the whole stuff.
The server might be down now and then today. Sorry.
January 5, 2008
The W6 workunits were too large. Therefore they have been canceled and replaced by 2688999
W6q workunits (q means quarter, every W6 has been replaced by 4 W6q). If you still have
W6 workunits you can cancel them or crunch them. You will get your credits.
January 3, 2008
Another bunch of workunits is ready: 764.000 workunits, named ''W6...'' are online
to feed your hungry machines. As we get closer to the project's target the problem
gets harder and harder. Therefore the maximum running time of the workunits is
10 hours now. If you have still W5's you can cancel them as their range of numbers
is included in the W6's. However, if you crunch them you will get your credits.
September 21, 2007
From November 14th 2006 until today we had no hardware problems at all. But today our
server wasn't able to write to the raid anymore. After a restart most services work
again and probably I will be able to reactivate the database and restart RCN today.
But there must be some hardware failure and therefore the crash might reoccur in the future.
Unfortunately I will be on a research stay in France from Sep. 25th until Oct. 8th.
So if the project isn't running, you might want to crunch workunits from our
Sudoku-Project (http://dist2.ist.tugraz.at/sudoku) meantime.
September 5, 2007
The announced SUDOKU project http://dist2.ist.tugraz.at/sudoku has started. Currently the
alpha tests are running (Linux only). Within the next view days applications for Windows
and Mac will be added. CU there.
July 9, 2007
Just to remind you: A power cut-off is scheduled for today and our server will go down in a few hours.
July 6, 2007
A logo contest for our upcoming Boinc project ''Sudoku'' has been started. See
the threads in our fora.
July 4, 2007
A power cut-off is scheduled for Moday, July 9th, 11:00 - 14:00 (GMT+1) and
the RCN server will be down during this time.
June 29, 2007
Great! More than 5 cpu years yesterday! And it seems to get even more
today. FYI: We work hard on the new Sudoko project. The new server
http://dist2.ist.tugraz.at/ is up an running. However, the scientific
application is not ready yet. I will announce the official start of the
project here. Best B.
June 23, 2007
Good good good news: 2 x Xeon 3.4 GHz, 5 x SCSI 146 GB, 4 GB RAM, ...
Yes, due to the successs of RCN we managed to get a second server.
It is already in my office and awaits its boincification. It will
be used as server for our upcoming project ''SUDOKU'' where we will
search for the smallest possible start configuration of the game.
The tests for the new project might start in one or two weeks. I
will keep you informed. Btw: Our users have established a new
record: Yesterday they contributed 3 cpu years!
June 20, 2007
Currently we have only a few tcape-crossing workunits left (W2,W2b,W2c) and
I have lowered the maximum buffer size to 2 workunits in order to achieve
the best possible distribution. We will start the W3 series very soon.
May 16, 2007
We have another 111 000 wednesday workunits. They are called W2 and they
are the same as the W1's but we have increased the maximum running time to
8 hours per workunit for all CPU types. Note that we have checkpointing now
and you won't lose your credit if you need to reboot your machine.
May 15, 2007
CHECKPOINTING in RCN: tcape-crossing has three new features now, namely
checkpoints, the progress bar is updated, and there is a tcape-crossing
version for Intel-Macs. Note that we still can't estimate the total
running time of a tcape-crossing workunit. Therefore the progress bar
shows the percent-done value regarding to the maximum running time of a
workunit. As a consequence there might be a jump from e.g. 10 % to 100 %
if a workunit is rather small. We've also tried to cross-compile the app
for PPC-Macs but it didn't work. We can't support PPCs without having a
PPC machine, sorry.
May 02, 2007
A first bunch of 177 000 wednesday workunits is online now. The names
start with W1... Much more wednesdays will follow as soon as the results
of the tuedrei's are back and the analysis is completed.
April 28, 2007
All remaining tuesday and TF workunits are online now. You should be able
to fetch them within the next 3 days.
March 26, 2007
We have no more tuedrei workunits left. Only 5400 are still in progress.
However, the analysis of your computed results shows that the problem is
still very hard and we have to generate another 213 000 tuesday workunits
before we can proceed with the wednesdays. We will call them TF (tuesday
final) and we will set them up with a shorter time limit of 2 hours (4 hours
for P4 HTs). Can you crunch them within the next 20 days? FYI: We will feed
your machines with 90 percent TF workunits and 10 percent rCross workunits
such that also your Intel-Macs will not starve. Btw: Have you already seen
the new system summary page? It contains a graph showing your cpu hours (the
maximum was 13870 hours!!! per day) and you get the information without
caching.
March 05, 2007
Mac news: I'm glad to announce that Intel-Macs are also supported now!
Today we have installed the new RCross application for Linux, Mac and
Windows-Users. It has checkpoints and it updates the progress bar.
Happy crunching...
February 22, 2007
Short maintenance downtime on Monday, 26. February 2007, 10:00. We
will activate database replication, reimplement our backup solution
and improve the summary script. If everything goes right you will
not notice anything at all.
January 18, 2007
We've started to test the additional scientific application rcross. It
has checkpoints and it updates the progress bar. We develop under Linux
and the first test version is only available for Linux crunchers because
it doesn't make sense to compile the first intermediate test versions
for all platforms. Currently we have about 10 workunits for rcross.
November 23, 2006
Short maintenance downtime. Just a few minutes.
November 21, 2006
Right now we are inserting 8400 decE and decF workunits. As soon as
they are solved we can do the final analysis regarding RCN's
first target.
November 14, 2006
A small hack of the server software allows us to prevent hosts
from accumulating unreasonably large workunit caches while other
users have nothing to crunch. As we have a rather small number of
workunits today we have set the maximum cache size to 10. Requests
of hosts having more than this number are rejected. As soon as we
have a sufficient number of workunits again we will increase this
value to 100 ore more. Please report if something doesn't work as
expected.
November 14, 2006
The new server works great! However, some computers have too
big caches and we don't expect to get results for thousands of
old workunits anymore. Therefore we ask everyone to cancel the
remaining tuesday, tuesdayR and tuesRetry workuntis. We will
also cancel them from the server and regenerate the missing
ones soon (see Number Crunching, I will not cancel anything
before thursday).
November 9, 2006
The new server has been delivered right now! The exchange will take
place within the next view days. I don't know when exactly. Just
don't worry if the project isn't reachable for a few hours.
November 6, 2006
The database is very slow. Due to reconfigurations we will have short
downtimes today and in the next few days. The good message is: We will
get a new server next week: Dual XEON, 4 GB, 5 SCSI disks :-) and our
service will be much better in the future.
November 1, 2006
Done. After some maintenance tasks the database runs better and we are online again.
November 1, 2006
Everything is slow. I have to stop the project for maintenance.
October 25, 2006
Update: We are up and crunching again.
October 25, 2006
Update: The recovery runs slowly. It will take several hours to restore
the database. We will go online as soon as possible.
October 25, 2006
One harddisk is defect. Recovery runs right now. Downtime for approximately
one hour. Sorry.
October 12, 2006
Great! Thanks to Asus' superior quality we had the chance to check the
quality of our backup strategy as well as the patience of our users.
Our mainboard died after only 4 months and the controllers messed up the
raid. Fortunately, we were able to restore everything from the backup.
Just the postings of one day (the ones after the last backup) are lost.
I had to reconfigure some parameters of the OS. If something doesn't
work, please use the fora. Btw, the server has 4 GB RAM now and this
should speed up the database :-) Sorry for the downtime!
October 2, 2006
This week we will shut down the server in order to attach additional
RAM modules. As the webserver runs on the same machine the web will
also not be available for a few minutes. Probably you will not notice
anything.
September 29, 2006
Longrunners this weekend: We have 40 000 workunits with a time limit of
4 hours instead of the usual 2 hours. However, most of them will terminate
earlier. You can identify these workunits by their name beginning with
'nov...'. Probably you will crunch them away over the weekend and afterwards
we will have the normal 2 hrs. time limit again.
September 27, 2006
The maintenance jobs are done. The server is much faster and right now
we are inserting a bunch of new workunits. Enough to crunch for all.
September 25, 2006
We are ready to implement some improvements today but we have to stop
the project for a few hours. We will also purge the database and the
server should be faster afterwards.
September 21, 2006
We have added a link to the server status page and we have a new
system summary page.
September 14, 2006
We are online again. See 'Message boards : Radio RCN'.
August 30, 2006
Please abort all RCN workunits now. Due to recent database trouble their
parameters are messy. You don't need to report them because I had to delete
them all. Don't worry about your credits. Generous credits will be given to
all users who had work in progress.
August 27, 2006
After some days of maintenance our server is online again. Everything should
work properly now.
August 25, 2006
We have some trouble with our server. See 'Message boards : Radio RCN'. Can we
fix it? Yes we can. I will try to do that tomorrow in the morning. Bernhard.
August 22, 2006
We will have some intervals of downtime today.
August 21, 2006
Maintenance Jobs: If there are still *old* workunits on your clients, please
abort them now. 'Old' means workunits that have to be computed with the old
'cape-crossing' application instead of the new 'tcape-crossing' application.
We will delete them from the database very soon and can not grant credits
afterwards. August 07, 2006
We have a new scientific application. It enforces a time limit for our workunits.
Currently the limit is set to 2 hours. But note that we measure the real cpu time.
For P4's with HT and two processes in parallel the maximum running time will be
doubled and also other processes at the computer may stretch the maximum running
time.
August 07, 2006
The winner of our logo contest is Cori. Congratulations, you did a great job!
August 01, 2006
Our users contributed some very good designs for a project logo. Now you can
vote for your favourite one, see 'Message boards : Radio RCN '.
And another thing: We started 4 weeks ago and now we have more than 1 000 participating
hosts. This is by far more than we expected and we are very proud of the project's
success. Time to say THANK YOU!
July 22, 2006
Logo contest officially started, see 'Message boards : Radio RCN : Project logo'.
In a few minutes we will have 10 000 new workunits and additionally I will insert
another big bunch of wu's tomorrow.
July 22, 2006
Accidentally a small number of workunits named 'july...' were released tonight.
Please abort them, they are all erroneous. Sorry for that. Now we have a new
series of workunits. They are named 'hotsummer...' and the name does not only
reflect the weather... Let's crunch!
July 18, 2006
We went dry of workunits today, but now there are new ones. Good luck and happy crunching!
July 15, 2006
The maintenance jobs are done. Our server is online again.
July 14, 2006
Remember, due to a power cut we will have some hours downtime tomorrow. This shouldn't
cause any problems for your clients and running workunits. Just for your information.
July 12, 2006
Another 30 000 workunits inserted. The time consumption is again unpredictable and might
be some seconds as well as several days (typically it is some hours). You've been warned.
Due to the unpredictable time behavior we can not update the progress bar.
July 10, 2006
The RCN-Community is pretty powerful: Only 29 000 WU's are left at the server. Time to
insert new ones. The newly inserted (you are likely to get them after 24h) are different
in that they are preselected such that no one takes less than 3 seconds. Unfortunately the
time consumption for a particular WU is unpredictable. Therefore we do not update the progress
bar. Running times of 5 seconds are as normal as running times of 150 hours, however the
latter should only occur in rare cases. I set the maximum time for the client's answer to
10 days. The measures will avoid the network overhead we encountered with
the xxx-short WU's.
July 10, 2006
*** As requested: From now on we export XML stats. ***
July 9, 2006
Here is my hotfix for the problem with the ultra-short WU's: Currently all machines of our
institute pre-process the remaining 322 000 WU's and kill the job if it takes more than
3 seconds. It seems that doing that we can _locally_ solve >70 percent until tomorrow. The
remaining longer WU's will be put to the server on monday and you are likely to get the new
WU-length-feeling on Thuesday or Wednesday when the old ones are finished.
July 7, 2006
Attention: We will have some hours downtime on July 15th because those guys will switch off
the power of our server room. The server will be up again as soon as possible. Hey, did I say thank
you? There are 49 new participants that joined us within the last two days. Currently we compute
over 16 000 workunits a day. I didn't reckon with this computing power and had to generate new
work earlier as estimated.
July 6, 2006
There are 60 000 workunits online waiting for your computing power. Since today we validate
the results and grant credits. All in all we have 550 000 workunits to compute. Most of them
will take just a view minutes but there are also some that will take several hours to complete.
June 30, 2006
We got a new server: AMD Athlon 64 X2 with 2 GB RAM and two 400 GB Harddisks. We have
also updated the software: Now the project is driven by the server edition of Ubuntu
Dapper Drake and Boinc V5.4. We are in the test phase again. Please inform us if
something goes wrong! Your Boinc admin.
