boinc/sched/update_stats.cpp

321 lines
9.3 KiB
C++
Raw Normal View History

// This file is part of BOINC.
// http://boinc.berkeley.edu
// Copyright (C) 2008 University of California
//
// BOINC is free software; you can redistribute it and/or modify it
// under the terms of the GNU Lesser General Public License
// as published by the Free Software Foundation,
// either version 3 of the License, or (at your option) any later version.
//
// BOINC is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
// See the GNU Lesser General Public License for more details.
//
// You should have received a copy of the GNU Lesser General Public License
// along with BOINC. If not, see <http://www.gnu.org/licenses/>.
// update_stats:
// Update average credit for idle users, hosts and teams.
// These fields are updates as new credit is granted;
// the purpose of this program is to decay credit of entities
// that are inactive for long periods.
// Run it about once a day.
//
// Also updates the nusers field of teams
//
// usage: update_stats args
// [--update_teams]
// [--update_users]
// [--update_hosts]
// [--min_age nsec] don't update items updated more recently than this
#include "config.h"
#include <cstdio>
#include <cstring>
#include <string>
#include <cstdlib>
#include <unistd.h>
#include "boinc_db.h"
#include "util.h"
#include "str_util.h"
#include "error_numbers.h"
#include "svn_version.h"
#include "sched_config.h"
#include "sched_util.h"
#include "sched_msgs.h"
// If the item's average credit has been updated more recently than this,
// don't update it (optimizes performance).
#define MIN_AGE 86400
double max_update_time;
int update_users() {
DB_USER user;
int retval;
char buf[256];
double now = dtime();
while (1) {
sprintf(buf, "where expavg_credit>0.1 and expavg_time < %f", max_update_time);
retval = user.enumerate(buf);
if (retval) {
if (retval != ERR_DB_NOT_FOUND) {
log_messages.printf(MSG_CRITICAL, "lost DB conn\n");
exit(1);
}
break;
}
update_average(
now, 0, 0, CREDIT_HALF_LIFE, user.expavg_credit, user.expavg_time
);
sprintf( buf, "expavg_credit=%f, expavg_time=%f",
user.expavg_credit, user.expavg_time
);
retval = user.update_field(buf);
if (retval) {
log_messages.printf(MSG_CRITICAL, "Can't update user %d\n", user.id);
return retval;
}
}
return 0;
}
int update_hosts() {
DB_HOST host;
int retval;
char buf[256];
double now = dtime();
while (1) {
sprintf(buf, "where expavg_credit>0.1 and expavg_time < %f", max_update_time);
retval = host.enumerate(buf);
if (retval) {
if (retval != ERR_DB_NOT_FOUND) {
log_messages.printf(MSG_CRITICAL, "lost DB conn\n");
exit(1);
}
break;
}
update_average(
now, 0, 0, CREDIT_HALF_LIFE, host.expavg_credit, host.expavg_time
);
sprintf(
buf,"expavg_credit=%f, expavg_time=%f",
host.expavg_credit, host.expavg_time
);
retval = host.update_field(buf);
if (retval) {
log_messages.printf(MSG_CRITICAL, "Can't update host %d\n", host.id);
return retval;
}
}
return 0;
}
int get_team_totals(TEAM& team) {
int nusers;
int retval;
DB_USER user;
char buf[256];
// count the number of users on the team
//
sprintf(buf, "where teamid=%d", team.id);
retval = user.count(nusers, buf);
if (retval) return retval;
if (team.nusers != nusers) {
log_messages.printf(MSG_CRITICAL,
"updating member count for [TEAM#%d]: database has %d users, count shows %d\n",
team.id, team.nusers, nusers
);
}
team.nusers = nusers;
return 0;
}
// fill in the nusers, total_credit and expavg_credit fields
// of the team table.
// This may take a while; don't do it often
//
int update_teams() {
DB_TEAM team;
int retval;
char buf[256];
double now = dtime();
while (1) {
retval = team.enumerate("where expavg_credit>0.1");
if (retval) {
if (retval != ERR_DB_NOT_FOUND) {
log_messages.printf(MSG_CRITICAL, "lost DB conn\n");
exit(1);
}
break;
}
retval = get_team_totals(team);
if (retval) {
log_messages.printf(MSG_CRITICAL,
"update_teams: get_team_credit([TEAM#%d]) failed: %d\n",
team.id,
retval
);
continue;
}
if (team.expavg_time < max_update_time) {
update_average(
now, 0, 0, CREDIT_HALF_LIFE, team.expavg_credit,
team.expavg_time
);
}
sprintf(
buf, "expavg_credit=%f, expavg_time=%f, nusers=%d",
team.expavg_credit, team.expavg_time, team.nusers
);
retval = team.update_field(buf);
if (retval) {
log_messages.printf(MSG_CRITICAL, "Can't update team %d\n", team.id);
return retval;
}
}
return 0;
}
void usage(char *name) {
fprintf(stderr,
"Update average credit for idle users, hosts and teams.\n"
"These fields are updates as new credit is granted;\n"
"the purpose of this program is to decay credit of entities\n"
"that are inactive for long periods.\n"
"Hence it should be run about once a day at most.\n\n"
"Also updates the nusers field of teams\n\n"
"Usage: %s [OPTION]...\n\n"
"Options:\n"
- server: change the following from per-host to per-(host, app version): - daily quota mechanism - reliable mechanism (accelerated retries) - "trusted" mechanism (adaptive replication) - scheduler: enforce host scale probation only for apps with host_scale_check set. - validator: do scale probation on invalid results (need this in addition to error and timeout cases) - feeder: update app version scales every 10 min, not 10 sec - back-end apps: support --foo as well as -foo for options Notes: - If you have, say, cuda, cuda23 and cuda_fermi plan classes, a host will have separate quotas for each one. That means it could error out on 100 jobs for cuda_fermi, and when its quota goes to zero, error out on 100 jobs for cuda23, etc. This is intentional; there may be cases where one version works but not the others. - host.error_rate and host.max_results_day are deprecated TODO: - the values in the app table for limits on jobs in progress etc. should override rather than config.xml. Implementation notes: scheduler: process_request(): read all host_app_versions for host at start; Compute "reliable" and "trusted" for each one. write modified records at end get_app_version(): add "reliable_only" arg; if set, use only reliable versions skip over-quota versions Multi-pass scheduling: if have at least one reliable version, do a pass for jobs that need reliable, and use only reliable versions. Then clear best_app_versions cache. Score-based scheduling: for need-reliable jobs, it will pick the fastest version, then give a score bonus if that version happens to be reliable. When get back a successful result from client: increase daily quota When get back an error result from client: impose scale probation decrease daily quota if not aborted Validator: when handling a WU, create a vector of HOST_APP_VERSION parallel to vector of RESULT. Pass it to assign_credit_set(). Make copies of originals so we can update only modified ones update HOST_APP_VERSION error rates Transitioner: decrease quota on timeout svn path=/trunk/boinc/; revision=21181
2010-04-15 03:13:56 +00:00
" [ -d X ] Set debug level to X\n"
" [ --update_teams ] Updates teams.\n"
" [ --update_users ] Updates users.\n"
" [ --update_hosts ] Updates hosts.\n"
" [ -h | --help ] Shows this help text\n"
" [ -v | --version ] Shows version information\n",
name
);
}
int main(int argc, char** argv) {
int retval, i;
bool do_update_teams = false;
bool do_update_users = false;
2007-04-18 20:49:58 +00:00
bool do_update_hosts = false;
max_update_time = time(0) - MIN_AGE;
check_stop_daemons();
for (i=1; i<argc; i++) {
- server: change the following from per-host to per-(host, app version): - daily quota mechanism - reliable mechanism (accelerated retries) - "trusted" mechanism (adaptive replication) - scheduler: enforce host scale probation only for apps with host_scale_check set. - validator: do scale probation on invalid results (need this in addition to error and timeout cases) - feeder: update app version scales every 10 min, not 10 sec - back-end apps: support --foo as well as -foo for options Notes: - If you have, say, cuda, cuda23 and cuda_fermi plan classes, a host will have separate quotas for each one. That means it could error out on 100 jobs for cuda_fermi, and when its quota goes to zero, error out on 100 jobs for cuda23, etc. This is intentional; there may be cases where one version works but not the others. - host.error_rate and host.max_results_day are deprecated TODO: - the values in the app table for limits on jobs in progress etc. should override rather than config.xml. Implementation notes: scheduler: process_request(): read all host_app_versions for host at start; Compute "reliable" and "trusted" for each one. write modified records at end get_app_version(): add "reliable_only" arg; if set, use only reliable versions skip over-quota versions Multi-pass scheduling: if have at least one reliable version, do a pass for jobs that need reliable, and use only reliable versions. Then clear best_app_versions cache. Score-based scheduling: for need-reliable jobs, it will pick the fastest version, then give a score bonus if that version happens to be reliable. When get back a successful result from client: increase daily quota When get back an error result from client: impose scale probation decrease daily quota if not aborted Validator: when handling a WU, create a vector of HOST_APP_VERSION parallel to vector of RESULT. Pass it to assign_credit_set(). Make copies of originals so we can update only modified ones update HOST_APP_VERSION error rates Transitioner: decrease quota on timeout svn path=/trunk/boinc/; revision=21181
2010-04-15 03:13:56 +00:00
if (is_arg(argv[i], "update_teams")) {
do_update_teams = true;
- server: change the following from per-host to per-(host, app version): - daily quota mechanism - reliable mechanism (accelerated retries) - "trusted" mechanism (adaptive replication) - scheduler: enforce host scale probation only for apps with host_scale_check set. - validator: do scale probation on invalid results (need this in addition to error and timeout cases) - feeder: update app version scales every 10 min, not 10 sec - back-end apps: support --foo as well as -foo for options Notes: - If you have, say, cuda, cuda23 and cuda_fermi plan classes, a host will have separate quotas for each one. That means it could error out on 100 jobs for cuda_fermi, and when its quota goes to zero, error out on 100 jobs for cuda23, etc. This is intentional; there may be cases where one version works but not the others. - host.error_rate and host.max_results_day are deprecated TODO: - the values in the app table for limits on jobs in progress etc. should override rather than config.xml. Implementation notes: scheduler: process_request(): read all host_app_versions for host at start; Compute "reliable" and "trusted" for each one. write modified records at end get_app_version(): add "reliable_only" arg; if set, use only reliable versions skip over-quota versions Multi-pass scheduling: if have at least one reliable version, do a pass for jobs that need reliable, and use only reliable versions. Then clear best_app_versions cache. Score-based scheduling: for need-reliable jobs, it will pick the fastest version, then give a score bonus if that version happens to be reliable. When get back a successful result from client: increase daily quota When get back an error result from client: impose scale probation decrease daily quota if not aborted Validator: when handling a WU, create a vector of HOST_APP_VERSION parallel to vector of RESULT. Pass it to assign_credit_set(). Make copies of originals so we can update only modified ones update HOST_APP_VERSION error rates Transitioner: decrease quota on timeout svn path=/trunk/boinc/; revision=21181
2010-04-15 03:13:56 +00:00
} else if (is_arg(argv[i], "update_users")) {
do_update_users = true;
- server: change the following from per-host to per-(host, app version): - daily quota mechanism - reliable mechanism (accelerated retries) - "trusted" mechanism (adaptive replication) - scheduler: enforce host scale probation only for apps with host_scale_check set. - validator: do scale probation on invalid results (need this in addition to error and timeout cases) - feeder: update app version scales every 10 min, not 10 sec - back-end apps: support --foo as well as -foo for options Notes: - If you have, say, cuda, cuda23 and cuda_fermi plan classes, a host will have separate quotas for each one. That means it could error out on 100 jobs for cuda_fermi, and when its quota goes to zero, error out on 100 jobs for cuda23, etc. This is intentional; there may be cases where one version works but not the others. - host.error_rate and host.max_results_day are deprecated TODO: - the values in the app table for limits on jobs in progress etc. should override rather than config.xml. Implementation notes: scheduler: process_request(): read all host_app_versions for host at start; Compute "reliable" and "trusted" for each one. write modified records at end get_app_version(): add "reliable_only" arg; if set, use only reliable versions skip over-quota versions Multi-pass scheduling: if have at least one reliable version, do a pass for jobs that need reliable, and use only reliable versions. Then clear best_app_versions cache. Score-based scheduling: for need-reliable jobs, it will pick the fastest version, then give a score bonus if that version happens to be reliable. When get back a successful result from client: increase daily quota When get back an error result from client: impose scale probation decrease daily quota if not aborted Validator: when handling a WU, create a vector of HOST_APP_VERSION parallel to vector of RESULT. Pass it to assign_credit_set(). Make copies of originals so we can update only modified ones update HOST_APP_VERSION error rates Transitioner: decrease quota on timeout svn path=/trunk/boinc/; revision=21181
2010-04-15 03:13:56 +00:00
} else if (is_arg(argv[i], "update_hosts")) {
do_update_hosts = true;
} else if (is_arg(argv[i], "min_age")) {
double x = atof(argv[++i]);
max_update_time = time(0) - x;
} else if (!strcmp(argv[i], "-d")) {
if (!argv[++i]) {
log_messages.printf(MSG_CRITICAL, "%s requires an argument\n\n", argv[--i]);
usage(argv[0]);
exit(1);
}
int dl = atoi(argv[i]);
log_messages.set_debug_level(dl);
if (dl == 4) g_print_queries = true;
- server: change the following from per-host to per-(host, app version): - daily quota mechanism - reliable mechanism (accelerated retries) - "trusted" mechanism (adaptive replication) - scheduler: enforce host scale probation only for apps with host_scale_check set. - validator: do scale probation on invalid results (need this in addition to error and timeout cases) - feeder: update app version scales every 10 min, not 10 sec - back-end apps: support --foo as well as -foo for options Notes: - If you have, say, cuda, cuda23 and cuda_fermi plan classes, a host will have separate quotas for each one. That means it could error out on 100 jobs for cuda_fermi, and when its quota goes to zero, error out on 100 jobs for cuda23, etc. This is intentional; there may be cases where one version works but not the others. - host.error_rate and host.max_results_day are deprecated TODO: - the values in the app table for limits on jobs in progress etc. should override rather than config.xml. Implementation notes: scheduler: process_request(): read all host_app_versions for host at start; Compute "reliable" and "trusted" for each one. write modified records at end get_app_version(): add "reliable_only" arg; if set, use only reliable versions skip over-quota versions Multi-pass scheduling: if have at least one reliable version, do a pass for jobs that need reliable, and use only reliable versions. Then clear best_app_versions cache. Score-based scheduling: for need-reliable jobs, it will pick the fastest version, then give a score bonus if that version happens to be reliable. When get back a successful result from client: increase daily quota When get back an error result from client: impose scale probation decrease daily quota if not aborted Validator: when handling a WU, create a vector of HOST_APP_VERSION parallel to vector of RESULT. Pass it to assign_credit_set(). Make copies of originals so we can update only modified ones update HOST_APP_VERSION error rates Transitioner: decrease quota on timeout svn path=/trunk/boinc/; revision=21181
2010-04-15 03:13:56 +00:00
} else if (!strcmp(argv[i], "-v") || !strcmp(argv[i], "--version")) {
printf("%s\n", SVN_VERSION);
exit(0);
- server: change the following from per-host to per-(host, app version): - daily quota mechanism - reliable mechanism (accelerated retries) - "trusted" mechanism (adaptive replication) - scheduler: enforce host scale probation only for apps with host_scale_check set. - validator: do scale probation on invalid results (need this in addition to error and timeout cases) - feeder: update app version scales every 10 min, not 10 sec - back-end apps: support --foo as well as -foo for options Notes: - If you have, say, cuda, cuda23 and cuda_fermi plan classes, a host will have separate quotas for each one. That means it could error out on 100 jobs for cuda_fermi, and when its quota goes to zero, error out on 100 jobs for cuda23, etc. This is intentional; there may be cases where one version works but not the others. - host.error_rate and host.max_results_day are deprecated TODO: - the values in the app table for limits on jobs in progress etc. should override rather than config.xml. Implementation notes: scheduler: process_request(): read all host_app_versions for host at start; Compute "reliable" and "trusted" for each one. write modified records at end get_app_version(): add "reliable_only" arg; if set, use only reliable versions skip over-quota versions Multi-pass scheduling: if have at least one reliable version, do a pass for jobs that need reliable, and use only reliable versions. Then clear best_app_versions cache. Score-based scheduling: for need-reliable jobs, it will pick the fastest version, then give a score bonus if that version happens to be reliable. When get back a successful result from client: increase daily quota When get back an error result from client: impose scale probation decrease daily quota if not aborted Validator: when handling a WU, create a vector of HOST_APP_VERSION parallel to vector of RESULT. Pass it to assign_credit_set(). Make copies of originals so we can update only modified ones update HOST_APP_VERSION error rates Transitioner: decrease quota on timeout svn path=/trunk/boinc/; revision=21181
2010-04-15 03:13:56 +00:00
} else if (!strcmp(argv[i], "-h") || !strcmp(argv[i], "--help")) {
usage(argv[0]);
exit(0);
} else {
log_messages.printf(MSG_CRITICAL, "unknown command line argument: %s\n\n", argv[i]);
usage(argv[0]);
exit(1);
}
}
// if no do_update flags set, set them all
//
if (!do_update_teams && !do_update_users && !do_update_hosts) {
do_update_teams = true;
do_update_users = true;
do_update_hosts = true;
}
log_messages.printf(MSG_NORMAL, "Starting\n");
retval = config.parse_file();
if (retval) {
log_messages.printf(MSG_CRITICAL,
"Can't parse config.xml: %s\n", boincerror(retval)
);
exit(1);
}
retval = boinc_db.open(config.db_name, config.db_host, config.db_user, config.db_passwd);
if (retval) {
log_messages.printf(MSG_CRITICAL, "Can't open DB\n");
exit(1);
}
retval = boinc_db.set_isolation_level(READ_UNCOMMITTED);
if (retval) {
log_messages.printf(MSG_CRITICAL,
"boinc_db.set_isolation_level: %s; %s\n",
boincerror(retval), boinc_db.error_string()
);
}
if (do_update_users) {
retval = update_users();
if (retval) {
log_messages.printf(MSG_CRITICAL,
"update_users failed: %s\n", boincerror(retval)
);
exit(1);
}
}
if (do_update_hosts) {
retval = update_hosts();
if (retval) {
log_messages.printf(MSG_CRITICAL,
"update_hosts failed: %s\n", boincerror(retval)
);
exit(1);
}
}
if (do_update_teams) {
retval = update_teams();
if (retval) {
log_messages.printf(MSG_CRITICAL,
"update_teams failed: %s\n", boincerror(retval)
);
exit(1);
}
}
log_messages.printf(MSG_NORMAL, "Finished\n");
return 0;
}
const char *BOINC_RCSID_6b05e9ecce = "$Id$";