Hello,
if someone is looking for the number of batches available, the info is in this forum/thread, first post
https://www.worldcommunitygrid.org/forums/wcg/viewthread_thread,42561

It says:
Current status of work units:

Available for Download: 8,479 batches
In Progress: 1,941 batches (15,362,727 work units)
Completed: 1,973 batches total - 1,973 batches in the last 30 days - average of 65 batches per day
Estimated backlog: 130 days

Just in case
All the best
MS

@daka said:

Man, I need to learn how to do this. I could set up a dedicated NFS share on my [future] FreeNAS/TrueNAS and take my headless game to the next level. I assume since all your clients PXE-boot and have NFS root filesystem that that eliminates the need for any local storage whatsoever? My clients use some combination of USB flash drive, spinning disk, or SSD, but as I build out a bigger farm, I need more ways to reduce management overhead. (I finally set up the unattended-upgrades Debian package that runs daily, so at least I'm pretty hands-off at the moment.)

I'm looking into learning configuration automation like Puppet/Chef/Ansible just to automate a lot of tedious BOINC stuff. Especially if I buy like 15+ Raspberry Pis or Odroids or something.

down to 4657 this morning

@hchc said:


Exactly, they have no local disk. 48 of them are bare boards from laptops with removed wifi-module, monitor and SSD, all mounted in two Plexiglas boxes (24 laptops in 350x360x250 mm space) with one power supply and a switch in each box.
The other 96 machines are temporary and still have the local disk in them, but it's not used.

The setup of PXE and NFS is surprisingly simple once you've done it a few times.
I have a separate vlan set up for it, where the DHCP-server (isc-dhcp-server) sends out the correct next-server and filename (depending on architecture: BIOS, efi64 or efi32). The TFTP-server (tftpd-hpa) serves pxelinux/syslinux with (currently) a separate config file per machine based on the system MAC-adress. The systems mount a folder per machine located under the NFS-exported folder (served by nfs-kernel-server).
It would be possible to make it even easier to manage but more complicated to set up by just using one dynamic configuration and have it load one default folder on each machine and then use a per machine overlay file where all the local changes are stored.

This installation is located in an upper secondary school and is primarily used for students who want to experiment with distributed calculations (for example generating terapixel sized images of Mandelbrot sets). In their idle time they contribute to WCG.

That was probably a bit off topic for this thread, and I haven't even got a new highest iteration to report... still on 4672.
Sorry about that

I think you are looking at the wrong numbers.