23:30 June 19, 2007 The database is back on-line. We have disabled the sending of new
work units and will be monitoring the system to see what happens.
Fortunately only one table of one database was corrupted.
Unfortunately it was the SciLINC results table. This will probably
result in no credit being granted for the results that are pending
being reported and we apologize for that.
More information will be posted tomorrow.
17:30 June 18, 2007 We have finished our examination and remote analysis of the main
SciLINC server. This lead us to believe that it was the victim of
electrical problems. It was later confirmed that there were some
transient electrical issues on Saturday in the building where the
machine is housed.
Binary junk was found written across a number of log files on this
server. Those files belong to various SciLINC processes, MySQL, Apache
and the Linux Kernel itself all around 7:02 PM local time Saturday the
16th. That's UTC 00:02 Sunday the 17th. This coincides with a
"disturbance in the force" that knocked a number of other machines
offline over the weekend as well.
These things should not happen, but they do. There are plans to
move the machine to a better location in the next couple of weeks.
At this point we are filing bug reports, as requested in the log
files, since MySQL is unable to recover from the transaction logs that
are present. As soon as this is done we will be attempting recovery
of the database and then work on getting the project back on-line.
We will keep you posted and thank you for your continued support.
11:05 June 18, 2007 New account creation has been temporarily disabled until we can iron
out some of the issues that we are having.
If you really must have an account or want to
help us test fixes to the CPU loading problem send us
an email requesting one and after things are back up we will try
to respond.
8:26 June 18, 2007 AlphaLaser and other users on the
BOINCstats forums and the
BOINC forums have brought to our attention an issue where the
large number of files being served up for each workunit are causing
high CPU loads in the core BOINC client.
Augustine found a partial workaround to this. It is
discussed on BOINCstats as well.
After the database is back on-line we will be looking into
correcting this issue.
8:00 June 18, 2007 We have experienced a database corruption. Most of the server
processes are down.
Updates will be posted here as they are available. At this time we
expect to have things back on-line within a day or two.
Thank you for your continued support and understanding.
June 15, 2007 The process that gathers new work for SciLINC got a little carried
away and filled the drive with everything it could find in the vaults.
Images, PDFs, everything. This has been corrected and the server
should now be running correctly again.
June 1, 2007 Results from SciLINC are now being passed to Botanicus.
May 18, 2007 We are having problems with the server. The scheduler, feeder and
associated processes are down, as you can see on the server status page.
May 14, 2007 We have moved SciLINC from our intranet to the Internet. The server
is now available via www.scilinc.org.
March 29, 2007 The SciLINC server sent out its first set of work units and received
its first results today, running on Linux. Expect a Windows version
shortly.
March 22, 2007 The SciLINC project went on-line today. We expect to be processing
work within a few days.
