boinc/sched/make_work.cpp

356 lines
11 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/>.
// make_work
- 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
// --wu_name name
// [ --wu_name name2 ... ]
// [ --cushion n ] // make work if fewer than N unsent results
// [ --max_wus n ] // don't make work if more than N total WUs
// [ --one_pass ] // quit after one pass
//
// Create WU and result records as needed to maintain a pool of work
// (for testing purposes).
// Clones the WU of the given name.
//
#include "config.h"
#include <cstdio>
#include <cstdlib>
#include <cstring>
#include <errno.h>
#include <unistd.h>
#include <ctime>
#include <vector>
#include <string>
using std::vector;
using std::string;
#include "boinc_db.h"
#include "crypt.h"
#include "util.h"
#include "backend_lib.h"
#include "sched_config.h"
#include "parse.h"
#include "sched_util.h"
#include "sched_msgs.h"
#include "str_util.h"
#include "svn_version.h"
#define LOCKFILE "make_work.out"
#define PIDFILE "make_work.pid"
int max_wus = 0;
int cushion = 300;
bool one_pass = false;
// edit a WU XML doc, replacing one filename by another
// (should appear twice, within <file_info> and <file_ref>)
// Don't patch the URL; we'll download the same file
//
void replace_file_name(char* xml_doc, char* filename, char* new_filename) {
char buf[BLOB_SIZE], temp[256];
char * p;
strcpy(buf, xml_doc);
p = strtok(buf,"\n");
while (p) {
if (parse_str(p, "<name>", temp, sizeof(temp))) {
- 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 (!strcmp(filename, temp)) {
replace_element_contents(
xml_doc + (p - buf),"<name>","</name>", new_filename
);
}
} else if (parse_str(p, "<file_name>", temp, sizeof(temp))) {
- 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 (!strcmp(filename, temp)) {
replace_element_contents(
xml_doc+(p-buf), "<file_name>","</file_name>", new_filename
);
}
}
p = strtok(0, "\n");
}
}
void make_new_wu(DB_WORKUNIT& original_wu, char* starting_xml, int start_time) {
char file_name[256], buf[BLOB_SIZE], new_file_name[256];
char new_buf[BLOB_SIZE];
char * p;
int retval;
DB_WORKUNIT wu = original_wu;
static int file_seqno = 0, wu_seqno = 0;
strcpy(buf, starting_xml);
p = strtok(buf, "\n");
strcpy(file_name, "");
// make new names for the WU's input files,
// so clients will download them.
// (don't actually copy files; URL stays the same)
//
while (p) {
if (parse_str(p, "<name>", file_name, sizeof(file_name))) {
sprintf(
new_file_name, "%s__%d_%d", file_name, start_time, file_seqno++
);
strcpy(new_buf, starting_xml);
replace_file_name(new_buf, file_name, new_file_name);
strcpy(wu.xml_doc, new_buf);
}
p = strtok(0, "\n");
}
// set various fields for new WU (all others are copied)
//
wu.id = 0;
wu.create_time = time(0);
// the name of the new WU cannot include the original WU name,
// because the original one probably contains "nodelete",
// but we want the copy to be eligible for file deletion
//
sprintf(wu.name, "wu_%d_%d", start_time, wu_seqno++);
wu.need_validate = false;
wu.canonical_resultid = 0;
wu.canonical_credit = 0;
wu.hr_class = 0;
wu.transition_time = time(0);
wu.error_mask = 0;
wu.file_delete_state = FILE_DELETE_INIT;
wu.assimilate_state = ASSIMILATE_INIT;
retval = wu.insert();
if (retval) {
log_messages.printf(MSG_CRITICAL,
"Failed to created WU: %s; exiting\n", boincerror(retval)
);
exit(retval);
}
original_wu.id = boinc_db.insert_id();
log_messages.printf(MSG_DEBUG,
"Created %s, clone of %s\n", wu.name, original_wu.name
);
}
// wait for the transitioner to create a result for the given WU.
// This keeps us from getting infinitely far ahead of the transitioner
// (e.g. if the transitioner isn't running)
//
void wait_for_results(int wu_id) {
DB_RESULT result;
int count, retval;
char buf[256];
sprintf(buf, "where workunitid=%d", wu_id);
while (1) {
retval = result.count(count, buf);
log_messages.printf(MSG_DEBUG, "result.count for %d returned %d, error: %s\n",
wu_id, count, boincerror(retval)
);
if (retval) {
log_messages.printf(MSG_CRITICAL, "result.count: %s\n", boincerror(retval));
exit(1);
}
if (count > 0) return;
sleep(10);
check_stop_daemons();
}
}
void make_work(vector<string> &wu_names) {
int retval, start_time=time(0);
char keypath[256];
char buf[BLOB_SIZE];
R_RSA_PRIVATE_KEY key;
int nwu_names = wu_names.size();
DB_WORKUNIT wus[nwu_names];
int i;
static int index=0;
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);
}
for (i=0; i<nwu_names; i++) {
DB_WORKUNIT& wu = wus[i];
sprintf(buf, "where name='%s'", wu_names[i].c_str());
retval = wu.lookup(buf);
if (retval) {
log_messages.printf(MSG_CRITICAL,
"can't find wu %s\n", wu_names[i].c_str()
);
exit(1);
}
}
sprintf(keypath, "%s/upload_private", config.key_dir);
retval = read_key_file(keypath, key);
if (retval) {
log_messages.printf(MSG_CRITICAL, "can't read key\n");
exit(1);
}
while (1) {
check_stop_daemons();
2007-04-18 20:49:58 +00:00
int unsent_results;
retval = count_unsent_results(unsent_results, wus[0].appid);
2007-04-18 20:49:58 +00:00
if (retval) {
log_messages.printf(MSG_CRITICAL,
2007-04-18 20:49:58 +00:00
"can't get result count\n"
);
exit(1);
}
int total_wus=0;
if (max_wus) {
retval = count_workunits(total_wus, "");
2007-04-18 20:49:58 +00:00
if (retval) {
log_messages.printf(MSG_CRITICAL,
2007-04-18 20:49:58 +00:00
"can't get wu count\n"
);
exit(1);
}
}
log_messages.printf(
MSG_DEBUG, "unsent: %d cushion: %d\n",
unsent_results, cushion
);
if (unsent_results > cushion) {
sleep(10);
continue;
}
int results_needed = cushion - unsent_results;
int new_wu_id = 0;
while (1) {
DB_WORKUNIT& wu = wus[index++];
if (index == nwu_names) index=0;
if (max_wus && total_wus >= max_wus) {
log_messages.printf(MSG_NORMAL,
"Reached max_wus = %d\n", max_wus
);
exit(0);
2007-04-18 20:49:58 +00:00
total_wus++;
}
make_new_wu(wu, wu.xml_doc, start_time);
new_wu_id = wu.id;
results_needed -= wu.target_nresults;
if (results_needed <= 0) break;
}
if (one_pass) break;
wait_for_results(new_wu_id);
}
}
void usage(char *name) {
fprintf(stderr,
"Create WU and result records as needed to maintain a pool of work\n"
"(for testing purposes).\n"
"Clones the WU of the given name.\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
" --wu_name name the name for the WU\n"
" (can be repeated)\n"
" [ --cushion N ] make work if fewer than N unsent results\n"
" [ --max_wus n ] don't make work if more than N total WUs\n"
" [ --one_pass ] quit after one pass\n"
" [ --d X ] set debug level to X.\n"
" [ -h | --help ] shows this help text.\n"
" [ -v | --version ] shows version information\n",
name
);
}
int main(int argc, char** argv) {
int i;
vector<string> wu_names;
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], "cushion")) {
if (!argv[++i]) {
log_messages.printf(MSG_CRITICAL, "%s requires an argument\n\n", argv[--i]);
usage(argv[0]);
exit(1);
}
cushion = atoi(argv[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
} else if (is_arg(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 (is_arg(argv[i], "wu_name")) {
if (!argv[++i]) {
log_messages.printf(MSG_CRITICAL, "%s requires an argument\n\n", argv[--i]);
usage(argv[0]);
exit(1);
}
wu_names.push_back(string(argv[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
} else if (is_arg(argv[i], "max_wus")) {
if (!argv[++i]) {
log_messages.printf(MSG_CRITICAL, "%s requires an argument\n\n", argv[--i]);
usage(argv[0]);
exit(1);
}
max_wus = atoi(argv[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
} else if (is_arg(argv[i], "one_pass")) {
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[1], "h") || is_arg(argv[1], "help")) {
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[1], "v") || is_arg(argv[1], "version")) {
printf("%s\n", SVN_VERSION);
exit(0);
} else {
log_messages.printf(MSG_CRITICAL, "unknown command line argument: %s\n\n", argv[i]);
usage(argv[0]);
exit(1);
}
}
check_stop_daemons();
if (!wu_names.size()) {
fprintf(stderr, "Must supply at least one WU name\n");
exit(1);
}
log_messages.printf(MSG_NORMAL,
"Starting: cushion %d, max_wus %d\n",
cushion, max_wus
);
install_stop_signal_handler();
srand48(getpid() + time(0));
make_work(wu_names);
}
const char *BOINC_RCSID_d24265dc7f = "$Id$";