2008-08-06 18:36:30 +00:00
|
|
|
// This file is part of BOINC.
|
2005-01-20 23:22:22 +00:00
|
|
|
// http://boinc.berkeley.edu
|
2008-08-06 18:36:30 +00:00
|
|
|
// Copyright (C) 2008 University of California
|
2004-07-10 00:19:10 +00:00
|
|
|
//
|
2008-08-06 18:36:30 +00:00
|
|
|
// 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.
|
2004-07-10 00:19:10 +00:00
|
|
|
//
|
2008-08-06 18:36:30 +00:00
|
|
|
// BOINC is distributed in the hope that it will be useful,
|
2005-01-20 23:22:22 +00:00
|
|
|
// 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.
|
2004-07-10 00:19:10 +00:00
|
|
|
//
|
2008-08-06 18:36:30 +00:00
|
|
|
// You should have received a copy of the GNU Lesser General Public License
|
|
|
|
// along with BOINC. If not, see <http://www.gnu.org/licenses/>.
|
2004-07-10 00:19:10 +00:00
|
|
|
|
|
|
|
//
|
|
|
|
// message_handler - check and validate new messages
|
- 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 debug_level]
|
|
|
|
// [--one_pass] // make one pass through table, then exit
|
2004-07-10 00:19:10 +00:00
|
|
|
//
|
|
|
|
// int handle_message(MSG_FROM_HOST&)
|
|
|
|
// handle a message from the host
|
|
|
|
//
|
|
|
|
// return nonzero on error
|
|
|
|
|
|
|
|
|
2005-11-21 18:34:44 +00:00
|
|
|
#include "config.h"
|
2004-07-10 00:19:10 +00:00
|
|
|
#include <unistd.h>
|
2008-02-27 23:26:38 +00:00
|
|
|
#include <cstdlib>
|
|
|
|
#include <string>
|
2004-07-10 00:19:10 +00:00
|
|
|
|
|
|
|
#include "boinc_db.h"
|
|
|
|
#include "util.h"
|
2008-03-27 18:25:29 +00:00
|
|
|
#include "error_numbers.h"
|
2007-05-11 16:30:13 +00:00
|
|
|
#include "str_util.h"
|
2009-09-17 17:56:59 +00:00
|
|
|
#include "svn_version.h"
|
2007-05-11 16:30:13 +00:00
|
|
|
|
2004-07-10 00:19:10 +00:00
|
|
|
#include "sched_config.h"
|
|
|
|
#include "sched_util.h"
|
|
|
|
#include "sched_msgs.h"
|
|
|
|
|
|
|
|
char app_name[256];
|
|
|
|
|
|
|
|
extern int handle_message(MSG_FROM_HOST&);
|
|
|
|
|
|
|
|
int handle_message(MSG_FROM_HOST& mfh) {
|
|
|
|
int retval;
|
|
|
|
|
2009-09-17 17:56:59 +00:00
|
|
|
printf("got message \n%s\n", mfh.xml);
|
2004-07-10 00:19:10 +00:00
|
|
|
DB_MSG_TO_HOST mth;
|
|
|
|
mth.clear();
|
|
|
|
mth.create_time = time(0);
|
|
|
|
mth.hostid = mfh.hostid;
|
|
|
|
mth.handled = false;
|
|
|
|
strcpy(mth.xml, mfh.xml);
|
|
|
|
retval = mth.insert();
|
|
|
|
if (retval) {
|
2010-11-08 17:51:57 +00:00
|
|
|
printf("insert failed %s\n", boincerror(retval));
|
2004-07-10 00:19:10 +00:00
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
// make one pass through msgs_from_host with handled == 0
|
|
|
|
// return true if there were any
|
|
|
|
//
|
|
|
|
bool do_message_scan() {
|
|
|
|
DB_MSG_FROM_HOST mfh;
|
|
|
|
char buf[256];
|
|
|
|
bool found=false;
|
|
|
|
int retval;
|
|
|
|
|
|
|
|
sprintf(buf, "where handled=0");
|
2008-03-27 18:25:29 +00:00
|
|
|
while (1) {
|
|
|
|
retval = mfh.enumerate(buf);
|
|
|
|
if (retval) {
|
|
|
|
if (retval != ERR_DB_NOT_FOUND) {
|
|
|
|
log_messages.printf(MSG_DEBUG,
|
|
|
|
"DB connection lost, exiting\n"
|
|
|
|
);
|
|
|
|
exit(0);
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
}
|
2004-07-10 00:19:10 +00:00
|
|
|
retval = handle_message(mfh);
|
|
|
|
if (!retval) {
|
|
|
|
mfh.handled = true;
|
|
|
|
mfh.update();
|
|
|
|
}
|
|
|
|
found = true;
|
|
|
|
}
|
|
|
|
return found;
|
|
|
|
}
|
|
|
|
|
|
|
|
int main_loop(bool one_pass) {
|
|
|
|
int retval;
|
|
|
|
bool did_something;
|
|
|
|
// char buf[256];
|
|
|
|
|
|
|
|
retval = boinc_db.open(config.db_name, config.db_host, config.db_user, config.db_passwd);
|
|
|
|
if (retval) {
|
2008-02-21 21:00:58 +00:00
|
|
|
log_messages.printf(MSG_CRITICAL,
|
2010-11-08 17:51:57 +00:00
|
|
|
"boinc_db.open failed: %s\n", boincerror(retval)
|
2005-09-26 23:28:48 +00:00
|
|
|
);
|
2004-07-10 00:19:10 +00:00
|
|
|
exit(1);
|
|
|
|
}
|
|
|
|
/*
|
|
|
|
sprintf(buf, "where name='%s'", app_name);
|
|
|
|
retval = app.lookup(buf);
|
|
|
|
if (retval) {
|
2008-02-21 21:00:58 +00:00
|
|
|
log_messages.printf(MSG_CRITICAL,
|
2005-09-26 23:28:48 +00:00
|
|
|
"can't find app %s\n", app_name
|
|
|
|
);
|
2004-07-10 00:19:10 +00:00
|
|
|
exit(1);
|
|
|
|
}
|
|
|
|
*/
|
|
|
|
while (1) {
|
|
|
|
check_stop_daemons();
|
|
|
|
did_something = do_message_scan();
|
|
|
|
if (one_pass) break;
|
|
|
|
if (!did_something) {
|
|
|
|
sleep(5);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2009-09-17 17:56:59 +00:00
|
|
|
void usage(char *name) {
|
|
|
|
fprintf(stderr,
|
|
|
|
"check and validate new messages\n\n"
|
|
|
|
"Usage: %s [OPTION]...\n\n"
|
|
|
|
"Options:\n"
|
|
|
|
" [ -d X ] Set debug level to X\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
|
|
|
" [ --one_pass ] make one pass through table, then exit\n"
|
|
|
|
" [ -h --help ] show this help text.\n"
|
|
|
|
" [ -v | --version ] show version information\n",
|
2009-09-17 17:56:59 +00:00
|
|
|
name
|
|
|
|
);
|
|
|
|
}
|
2004-07-10 00:19:10 +00:00
|
|
|
|
|
|
|
int main(int argc, char** argv) {
|
|
|
|
int i, retval;
|
2007-04-18 20:49:58 +00:00
|
|
|
bool one_pass = false;
|
2004-07-10 00:19:10 +00:00
|
|
|
|
|
|
|
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], "one_pass")) {
|
2004-07-10 00:19:10 +00:00
|
|
|
one_pass = 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], "d")) {
|
|
|
|
if (!argv[++i]) {
|
2009-09-17 17:56:59 +00:00
|
|
|
log_messages.printf(MSG_CRITICAL, "%s requires an argument\n\n", argv[--i]);
|
|
|
|
usage(argv[0]);
|
|
|
|
exit(1);
|
|
|
|
}
|
2010-04-05 21:59:33 +00:00
|
|
|
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 (is_arg(argv[i], "h") || is_arg(argv[i], "help")) {
|
2009-09-17 17:56:59 +00:00
|
|
|
usage(argv[0]);
|
|
|
|
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 (is_arg(argv[i], "v") || is_arg(argv[i], "version")) {
|
2009-09-17 17:56:59 +00:00
|
|
|
printf("%s\n", SVN_VERSION);
|
|
|
|
exit(0);
|
2004-07-10 00:19:10 +00:00
|
|
|
} else {
|
2009-09-17 17:56:59 +00:00
|
|
|
log_messages.printf(MSG_CRITICAL, "unknown command line argument: %s\n\n", argv[i]);
|
|
|
|
usage(argv[0]);
|
|
|
|
exit(1);
|
2004-07-10 00:19:10 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2009-05-07 13:54:51 +00:00
|
|
|
retval = config.parse_file();
|
2004-07-10 00:19:10 +00:00
|
|
|
if (retval) {
|
2008-02-21 21:00:58 +00:00
|
|
|
log_messages.printf(MSG_CRITICAL,
|
2009-05-07 13:54:51 +00:00
|
|
|
"Can't parse config.xml: %s\n", boincerror(retval)
|
2004-07-10 00:19:10 +00:00
|
|
|
);
|
|
|
|
exit(1);
|
|
|
|
}
|
|
|
|
|
2008-02-21 21:00:58 +00:00
|
|
|
log_messages.printf(MSG_NORMAL, "Starting message handler\n");
|
2004-07-10 00:19:10 +00:00
|
|
|
|
|
|
|
install_stop_signal_handler();
|
|
|
|
|
|
|
|
main_loop(one_pass);
|
|
|
|
}
|
2004-12-08 00:40:19 +00:00
|
|
|
|
2005-01-02 18:29:53 +00:00
|
|
|
const char *BOINC_RCSID_ff3b9880d4 = "$Id$";
|