PrimeGrid administrator responsibilities

These are (hopefully) all the things we do to keep PrimeGrid running:
 * Server Administration (Tyler) ( walli ) (ILW8)
 * Setting up new Linux servers
 * Server security
 * Monitoring servers for problems
 * backups
 * database administration
 * database replication
 * rsync from primary to backup system
 * making sure we don't run out of disk space (mostly making sure logrotate does what it's supposed to do)
 * OKing new profiles and profile pics. Nothing racy or controversial permitted


 * Loading work ( walli, Reggie? ) ( Mike Gutierrez )
 * The system sends out an email when it hits a preset minimum number of workunits remaining
 * A work file is a section of a sieve file for the n range above what was last loaded
 * Unless a subproject is being actively sieved, work files can be produced in advance of when needed
 * Managing the sieving process (Tyler) ( walli ) (ILW8)
 * producing work files from the manual sieve
 * running manual sieving
 * download new sieving at least once a day
 * check that sieving for errors (gaps, early ending, wrong n value, wrong p values)
 * PM users where sieving is unacceptable, remove it from the server so they can upload again
 * create an up-to-date sieve file
 * OK the ranges on the sieving system to give out credit. Credit is distributed at 0400 UTC
 * Calculate the optimal sieving level for work done over the next 5 years
 * Update sieving stats and any graphs involved
 * Setting up BOINC sieving when appropriate
 * Starting new sieving for new n ranges before running on on PrimeGrid
 * Every three months, resieve PPS/RSP 6M-9M and make the RSP sieves available to everyone. Update the thread on mersenneforum to let everyone know.
 * Loading new work into PRPNet when needed (Reggie?) ( Mike Gutierrez )
 * These candidates come from the PPS/RSP sieving. Because of the DIV project we no longer load k=27 candidates with c=1 as they're tested in DIV
 * Occasionally take residues from PRPNet and load them into BOINC so that only one new task is sent out
 * Checking to see if new primes were found on PRPNet (Reggie) ( Mike Gutierrez )
 * Managing BOINC apps ( walli )
 * Mike's app generator tool
 * new projects
 * new software versions for existing projects
 * detecting and deleting old app versions that are no longer used
 * Developing apps ( Mike Gutierrez )
 * Overall process run by the Software Coordinator (currently stream)
 * apps developed by various developers
 * rigorous acceptance testing (culminating in public testing on the dev server)
 * Watch out for the unwashed hordes gobbling up all the tasks on dev trying to get WuProp hours)
 * Assigning manual credit for PRPNet once a week (or automating the process) (Reggie) ( Mike Gutierrez )


 * Challenges ( walli ) ( Mike Gutierrez )


 * Setting up the server's challenge system configuration for each challenge
 * Making tweeks to BOINC server so it doesn't blow up on the more difficult challenges
 * Everything the Challenge Coordinator does (that's Mike Gutierrez): creating the challenge schedule, being the public face of the challenges.
 * TdP:
 * Setting up the special TdP badges.  New badges need to be created starting in 2021.  Don't forget to add them to the zip file for the 3rd part stats and badge sites.
 * Modifying front page to show TdP information (uncomment a few lines)
 * When all the primes are finalized on the server, remove the "red" warnings on the TdP page.
 * Reporting new primes (Reggie) ( Mike Gutierrez )
 * Primes large enough to be reportable will not show up until T5K verifies them
 * Most prime reporting involves one click and a copy/paste
 * Occasionally a new proof code must be added. Again, that's a single click, copy & paste.
 * If a user has not given permission to report primes, he/she must be contacted
 * Create official announcements. Currently done by Scott Brown, but we could use a backup. (Reggie)
 * Updating forum posts when significant primes are found. (Reggie)
 * Updating the front page when significant primes are found. (Reggie)
 * Providing an off-site backup location. The two servers both have a copy of EVERYTHING, but a site disaster could kill it all.  Jim's house currently serves as the off-site backup location.  We'll need a off-site backup. (Tyler) ( walli ) (ILW8)
 * Maintaining ALL the BOINC web pages. Creating new stuff.  Not killing the database performance while creating cool new stuff. ( walli ) (ILW8)
 * Fundraising, when necessary. It will be necessary this year (2020). [Unless hardware fails, we have funds for 24 months - ~Rytis] [Rytis]
 * Understanding how our server SSL certificates work. Renewal is automatic through "Let's Encrypt".  No action should be needed unless a new server is set up, or something breaks.
 * Controlling (granting) access to exported statistics, i.e., the www.primegrid.com/stats/ directory.
 * Running benchmarks for new FFT sizes (Pavel)