Since a few days the download speed of SETI XML files is really slow (at least to our servers), hovering between 5~25kb/s. Due to this, downloading the SETI host file takes over five (5!) hours. User XML takes over one hour.

Until the download speed is back to normal (meaning, somewhere nearer to the 1.5Mb/s we get with most other projects) or an other solution is found (thinking about downloading the files out of sync with the updates) the SETI host file will not be download, meaning no updates for SETI host stats.

I'm sorry for the inconvenience, but delaying updates this long is unacceptable.

The user XML will be downloaded and processed (unless they drop to 1kb/s).

David just informed me that SETI sysadmins are working with the RAID configurations. Should be ok withing a day or two.

Today's host update crashed because the SETI download took to long (the mysql connection of the update script was terminated). Because a lock prevents that two updates start at the same time, and the host update didn't remove the lock because it crashed, the daily update didn't start. I was busy working on the new PG server (damn Debian for not supporting the hardware) and I didn't notice it until 15 minutes ago.

It's now running (that is, starting with a one hour download of the SETI user XML file) so it's coming, but it's late.

I thought SETI's problems were over, boy was I wrong. No more SETI host updates until it is confirmed that the download speed is normal again.

really getting fed up with this shit stopping the stats over and over again.

Ditto.

Look at the brightside:

It gives me an excuse to post a link to the forum to remind the lurkers that BOINCstats is more than a shoutbox. . . .

I decoupled the XML downloads and the updates (meaning they can run in parallel now). This should solve the long delays although it may slow down the updates a little bit when XML is imported during a stats update (giving MySQL just a little bit more to do).