Rooba
/
boinc
mirror of https://github.com/BOINC/boinc.git
1
0
Fork 0
Code Issues 1 Packages Projects Releases Wiki Activity
boinc/sched/handle_request.cpp

1528 lines
52 KiB
C++
Raw Blame 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/>.
// Handle a scheduling server RPC
#include "config.h"
#ifdef _USING_FCGI_
#include "boinc_fcgi.h"
#else
#include <cstdio>
#endif
#include <cassert>
#include <cstdlib>
#include <vector>
#include <string>
#include <cstring>
#include <ctime>
#include <cmath>
#include <unistd.h>
#include <sys/wait.h>
#include <unistd.h>
#include <fcntl.h>
#include <errno.h>
#include <sys/stat.h>
#include "boinc_db.h"
#include "backend_lib.h"
#include "error_numbers.h"
#include "parse.h"
#include "str_util.h"
#include "str_replace.h"
#include "util.h"
#include "filesys.h"
#include "main.h"
#include "server_types.h"
#include "sched_util.h"
#include "handle_request.h"
#include "sched_msgs.h"
#include "sched_resend.h"
#include "sched_send.h"
#include "sched_config.h"
#include "sched_locality.h"
#include "sched_result.h"
#include "sched_plan.h"
#include "time_stats_log.h"
// find the user's most recently-created host with given various characteristics
//
static bool find_host_by_other(DB_USER& user, HOST req_host, DB_HOST& host) {
char buf[2048];
char dn[512], ip[512], os[512], pm[512];
#ifdef EINSTEIN_AT_HOME
// This is to prevent GRID hosts that manipulate their hostids from flooding E@H's DB with slow queries
if ((user.id == 282952) || (user.id == 243543))
return false;
#endif
// Only check if the fields are populated
if (strlen(req_host.domain_name) && strlen(req_host.last_ip_addr) && strlen(req_host.os_name) && strlen(req_host.p_model)) {
strcpy(dn, req_host.domain_name);
escape_string(dn, 512);
strcpy(ip, req_host.last_ip_addr);
escape_string(ip, 512);
strcpy(os, req_host.os_name);
escape_string(os, 512);
strcpy(pm, req_host.p_model);
escape_string(pm, 512);
sprintf(buf,
"where userid=%d and id>%d and domain_name='%s' and last_ip_addr = '%s' and os_name = '%s' and p_model = '%s'"
" and m_nbytes = %lf order by id desc", user.id, req_host.id, dn, ip, os, pm, req_host.m_nbytes
);
if (!host.enumerate(buf)) {
host.end_enumerate();
return true;
}
}
return false;
}
static void get_weak_auth(USER& user, char* buf) {
char buf2[256], out[256];
sprintf(buf2, "%s%s", user.authenticator, user.passwd_hash);
md5_block((unsigned char*)buf2, strlen(buf2), out);
sprintf(buf, "%d_%s", user.id, out);
}
static void send_error_message(const char* msg, int delay) {
g_reply->insert_message(USER_MESSAGE(msg, "low"));
g_reply->set_delay(delay);
g_reply->nucleus_only = true;
}
// Try to lock a file with name based on host ID,
// to prevent 2 schedulers from running at same time for same host.
// Return:
// 0 if successful
// In this case store file descriptor in reply struct so we can unlock later
// In other cases store -1 in reply struct
// PID (>0) if another process has lock
// -1 if error (e.g. can't create file)
//
int lock_sched() {
char filename[256];
char pid_string[16];
int fd, pid, count;
g_reply->lockfile_fd=-1;
sprintf(filename, "%s/CGI_%07d", config.sched_lockfile_dir, g_reply->host.id);
fd = open(filename, O_WRONLY|O_CREAT, S_IRUSR|S_IWUSR|S_IRGRP|S_IROTH);
if (fd < 0) return -1;
// if we can't get an advisory write lock on the file,
// return the PID of the process that DOES hold the lock.
// (or -1 if failure)
//
pid = mylockf(fd);
if (pid) {
close(fd);
return pid;
}
// write PID into the CGI_<HOSTID> file and flush to disk
//
count = sprintf(pid_string, "%d\n", getpid());
write(fd, pid_string, count);
fsync(fd);
g_reply->lockfile_fd = fd;
return 0;
}
// unlock and delete per-host lockfile
//
void unlock_sched() {
char filename[256];
if (g_reply->lockfile_fd < 0) return;
sprintf(filename, "%s/CGI_%07d", config.sched_lockfile_dir, g_reply->host.id);
unlink(filename);
close(g_reply->lockfile_fd);
}
// find the user's most recently-created host with given host CPID
//
static bool find_host_by_cpid(DB_USER& user, char* host_cpid, DB_HOST& host) {
char buf[256], buf2[256];
sprintf(buf, "%s%s", host_cpid, user.email_addr);
md5_block((const unsigned char*)buf, strlen(buf), buf2);
sprintf(buf,
"where userid=%d and host_cpid='%s' order by id desc", user.id, buf2
);
if (!host.enumerate(buf)) {
host.end_enumerate();
return true;
}
return false;
}
// Called when there's evidence that the host has detached.
// Mark in-progress results for the given host
// as server state OVER, outcome CLIENT_DETACHED.
// This serves two purposes:
// 1) make sure we don't resend these results to the host
// (they may be the reason the user detached)
// 2) trigger the generation of new results for these WUs
//
static void mark_results_over(DB_HOST& host) {
char buf[256], buf2[256];
DB_RESULT result;
sprintf(buf, "where hostid=%d and server_state=%d",
host.id,
RESULT_SERVER_STATE_IN_PROGRESS
);
while (!result.enumerate(buf)) {
sprintf(buf2,
"server_state=%d, outcome=%d, received_time = %ld",
RESULT_SERVER_STATE_OVER,
RESULT_OUTCOME_CLIENT_DETACHED,
time(0)
);
result.update_field(buf2);
// and trigger WU transition
//
DB_WORKUNIT wu;
wu.id = result.workunitid;
sprintf(buf2, "transition_time=%d", (int)time(0));
wu.update_field(buf2);
log_messages.printf(MSG_CRITICAL,
"[HOST#%d] [RESULT#%d] [WU#%d] changed CPID: marking in-progress result %s as client error!\n",
host.id, result.id, result.workunitid, result.name
);
}
}
// Based on the info in the request message,
// look up the host and its user, and make sure the authenticator matches.
// Some special cases:
// 1) If no host ID is supplied, or if RPC seqno mismatch,
// create a new host record
// 2) If the host record specified by g_request->hostid is a "zombie"
// (i.e. it was merged with another host via the web site)
// then follow links to find the proper host
//
// POSTCONDITION:
// If this function returns zero, then:
// - reply.host contains a valid host record (possibly new)
// - reply.user contains a valid user record
// - if user belongs to a team, reply.team contains team record
//
int authenticate_user() {
int retval;
char buf[256];
DB_HOST host;
DB_USER user;
DB_TEAM team;
if (g_request->hostid) {
retval = host.lookup_id(g_request->hostid);
while (!retval && host.userid==0) {
// if host record is zombie, follow link to new host
//
retval = host.lookup_id(host.rpc_seqno);
if (!retval) {
g_reply->hostid = host.id;
log_messages.printf(MSG_NORMAL,
"[HOST#%d] forwarding to new host ID %d\n",
g_request->hostid, host.id
);
}
}
if (retval) {
g_reply->insert_message(
USER_MESSAGE("Can't find host record", "low")
);
log_messages.printf(MSG_NORMAL,
"[HOST#%d?] can't find host\n",
g_request->hostid
);
g_request->hostid = 0;
goto lookup_user_and_make_new_host;
}
g_reply->host = host;
// look up user based on the ID in host record,
// and see if the authenticator matches (regular or weak)
//
sprintf(buf, "where id=%d", host.userid);
retval = user.lookup(buf);
if (!retval && !strcmp(user.authenticator, g_request->authenticator)) {
// req auth matches user auth - go on
} else {
bool weak_auth = false;
if (!retval) {
// user for host.userid exists - check weak auth
//
get_weak_auth(user, buf);
if (!strcmp(buf, g_request->authenticator)) {
weak_auth = true;
log_messages.printf(MSG_DEBUG,
"[HOST#%d] accepting weak authenticator\n",
host.id
);
}
}
if (!weak_auth) {
// weak auth failed - look up user based on authenticator
//
strlcpy(
user.authenticator, g_request->authenticator, sizeof(user.authenticator)
);
sprintf(buf, "where authenticator='%s'", user.authenticator);
retval = user.lookup(buf);
if (retval) {
g_reply->insert_message(
USER_MESSAGE("Invalid or missing account key. "
"Detach and reattach to this project to fix this.",
"high"
)
);
g_reply->set_delay(DELAY_MISSING_KEY);
g_reply->nucleus_only = true;
log_messages.printf(MSG_CRITICAL,
"[HOST#%d] [USER#%d] Bad authenticator '%s'\n",
host.id, user.id, g_request->authenticator
);
return ERR_AUTHENTICATOR;
}
}
}
g_reply->user = user;
if (host.userid != user.id) {
// If the request's host ID isn't consistent with the authenticator,
// create a new host record.
//
log_messages.printf(MSG_NORMAL,
"[HOST#%d] [USER#%d] inconsistent host ID; creating new host\n",
host.id, user.id
);
goto make_new_host;
}
// If the seqno from the host is less than what we expect,
// the user must have copied the state file to a different host.
// Make a new host record.
//
if (!batch && g_request->rpc_seqno < g_reply->host.rpc_seqno) {
g_request->hostid = 0;
log_messages.printf(MSG_NORMAL,
"[HOST#%d] [USER#%d] RPC seqno %d less than expected %d; creating new host\n",
g_reply->host.id, user.id, g_request->rpc_seqno, g_reply->host.rpc_seqno
);
goto make_new_host;
}
} else {
// Here no hostid was given, or the ID was bad.
// Look up the user, then create a new host record
//
lookup_user_and_make_new_host:
// if authenticator contains _, it's a weak auth
//
if (strchr(g_request->authenticator, '_')) {
int userid = atoi(g_request->authenticator);
retval = user.lookup_id(userid);
if (!retval) {
get_weak_auth(user, buf);
if (strcmp(buf, g_request->authenticator)) {
retval = ERR_AUTHENTICATOR;
}
}
} else {
strlcpy(
user.authenticator, g_request->authenticator,
sizeof(user.authenticator)
);
sprintf(buf, "where authenticator='%s'", user.authenticator);
retval = user.lookup(buf);
}
if (retval) {
g_reply->insert_message(
USER_MESSAGE(
"Invalid or missing account key. "
"Detach and reattach to this project to fix this.",
"low"
)
);
g_reply->set_delay(DELAY_MISSING_KEY);
log_messages.printf(MSG_CRITICAL,
"[HOST#<none>] Bad authenticator '%s': %d\n",
g_request->authenticator, retval
);
return ERR_AUTHENTICATOR;
}
g_reply->user = user;
// If host CPID is present,
// scan backwards through this user's hosts,
// looking for one with the same host CPID.
// If we find one, it means the user detached and reattached.
// Use the existing host record,
// and mark in-progress results as over.
//
if (strlen(g_request->host.host_cpid)) {
if (find_host_by_cpid(user, g_request->host.host_cpid, host)) {
log_messages.printf(MSG_CRITICAL,
"[HOST#%d] [USER#%d] User has another host with same CPID.\n",
host.id, host.userid
);
if (!config.multiple_clients_per_host) {
mark_results_over(host);
}
goto got_host;
}
}
make_new_host:
// One final attempt to locate an existing host record:
// scan backwards through this user's hosts,
// looking for one with the same host name,
// IP address, processor and amount of RAM.
// If found, use the existing host record,
// and mark in-progress results as over.
//
// NOTE: if the project allows multiple clients per host
// (e.g. those that run on grids), skip this.
//
if (!config.multiple_clients_per_host && find_host_by_other(user, g_request->host, host)) {
log_messages.printf(MSG_NORMAL,
"[HOST#%d] [USER#%d] Found similar existing host for this user - assigned.\n",
host.id, host.userid
);
mark_results_over(host);
goto got_host;
}
// either of the above cases,
// or host ID didn't match user ID,
// or RPC seqno was too low.
//
// Create a new host.
// g_reply->user is filled in and valid at this point
//
host = g_request->host;
host.id = 0;
host.create_time = time(0);
host.userid = g_reply->user.id;
host.rpc_seqno = 0;
host.expavg_time = time(0);
host.error_rate = 0.1;
strcpy(host.venue, g_reply->user.venue);
host.fix_nans();
retval = host.insert();
if (retval) {
g_reply->insert_message(
USER_MESSAGE("Couldn't create host record in database", "low")
);
boinc_db.print_error("host.insert()");
log_messages.printf(MSG_CRITICAL, "host.insert() failed\n");
return retval;
}
host.id = boinc_db.insert_id();
got_host:
g_reply->host = host;
g_reply->hostid = g_reply->host.id;
// this tells client to updates its host ID
g_request->rpc_seqno = 0;
// this value eventually gets written to host DB record;
// for new hosts it must be zero.
// This kludge forces this.
}
// have user record in g_reply->user at this point
//
if (g_reply->user.teamid) {
retval = team.lookup_id(g_reply->user.teamid);
if (!retval) g_reply->team = team;
}
// compute email hash
//
md5_block(
(unsigned char*)g_reply->user.email_addr,
strlen(g_reply->user.email_addr),
g_reply->email_hash
);
// if new user CPID, update user record
//
if (strlen(g_request->cross_project_id)) {
if (strcmp(g_request->cross_project_id, g_reply->user.cross_project_id)) {
user.id = g_reply->user.id;
escape_string(g_request->cross_project_id, sizeof(g_request->cross_project_id));
sprintf(buf, "cross_project_id='%s'", g_request->cross_project_id);
unescape_string(g_request->cross_project_id, sizeof(g_request->cross_project_id));
user.update_field(buf);
}
}
return 0;
}
// Modify claimed credit based on the historical granted credit if
// the project is configured to do this
//
static void modify_credit_rating(HOST& host) {
double new_claimed_credit = 0;
double percent_difference = 0;
// The percent difference between claim and history
double difference_weight = 1;
// The weight to be applied based on the difference between claim and
// history
double credit_weight = 1;
// The weight to be applied based on how much credit the host has earned
// (hosts that are new do not have accurate histories so they shouldn't
// have much weight)
double combined_weight = 1;
// Only modify if the credit_per_cpu_sec is established
// and the option is enabled
if ( host.credit_per_cpu_sec > 0
&& config.granted_credit_weight > 0.0
&& config.granted_credit_weight <= 1.0 ) {
// Calculate the difference between claimed credit and the hosts
// historical granted credit history
percent_difference=host.claimed_credit_per_cpu_sec-host.credit_per_cpu_sec;
percent_difference = fabs(percent_difference/host.credit_per_cpu_sec);
// A study on World Community Grid determined that 50% of hosts
// claimed within 10% of their historical credit per cpu sec.
// These hosts should not have their credit modified.
if ( percent_difference < 0.1000 ) {
log_messages.printf(MSG_DEBUG, "[HOSTID:%d] Claimed credit %.1lf not "
"modified. Percent Difference %.4lf\n", host.id,
host.claimed_credit_per_cpu_sec*86400, percent_difference
);
return;
}
// The study also determined that 95% of hosts claim within
// 50% of their historical credit per cpu sec.
// Computers claiming above 10% but below 50% should have their
// credit adjusted based on their history
// Computers claiming more than 50% above should use their
// historical value.
if ( percent_difference < .5 ) {
// weight based on variance from historical credit
difference_weight=1-(0.5-percent_difference)/0.4;
} else {
difference_weight=1;
}
// A weight also needs to be calculated based upon the amount of
// credit awarded to a host. This is becuase hosts without much
// credit awarded do not yet have an accurate history so the weight
// should be limited for these hosts.
if ( config.granted_credit_ramp_up ) {
credit_weight=config.granted_credit_ramp_up - host.total_credit;
credit_weight=credit_weight/config.granted_credit_ramp_up;
if ( credit_weight < 0) credit_weight = 0;
credit_weight = 1 - credit_weight;
}
// Compute the combined weight
combined_weight=credit_weight*difference_weight*config.granted_credit_weight;
log_messages.printf(MSG_DEBUG, "[HOSTID:%d] Weight details: "
"diff_weight=%.4lf credit_weight=%.4lf config_weight=%.4lf\n",
host.id, difference_weight, credit_weight,
config.granted_credit_weight
);
// Compute the new value for claimed credit
new_claimed_credit=(1-combined_weight)*host.claimed_credit_per_cpu_sec;
new_claimed_credit=new_claimed_credit+combined_weight*host.credit_per_cpu_sec;
if ( new_claimed_credit < host.claimed_credit_per_cpu_sec ) {
log_messages.printf(MSG_DEBUG, "[HOSTID:%d] Modified claimed credit "
"(lowered) original: %.1lf new: %.1lf historical: %.1lf "
"combined weight: %.4lf\n", host.id,
host.claimed_credit_per_cpu_sec*86400,
new_claimed_credit*86400, host.credit_per_cpu_sec*86400,
combined_weight
);
} else {
log_messages.printf(MSG_DEBUG, "[HOSTID:%d] Modified claimed credit "
"(increased) original: %.1lf new: %.1lf historical: %.1lf "
"combined weight: %.4lf\n", host.id,
host.claimed_credit_per_cpu_sec*86400,
new_claimed_credit*86400, host.credit_per_cpu_sec*86400,
combined_weight
);
}
host.claimed_credit_per_cpu_sec=new_claimed_credit;
}
}
// somewhat arbitrary formula for credit as a function of CPU time.
// Could also include terms for RAM size, network speed etc.
//
static void compute_credit_rating(HOST& host) {
double fpw, intw, scale, x;
if (config.use_benchmark_weights) {
fpw = config.fp_benchmark_weight;
intw = 1. - fpw;
// FP benchmark is 2x int benchmark, on average.
// Compute a scaling factor the gives the same credit per day
// no matter how benchmarks are weighted
//
scale = 1.5 / (2*intw + fpw);
} else {
fpw = .5;
intw = .5;
scale = 1;
}
x = fpw*fabs(host.p_fpops) + intw*fabs(host.p_iops);
x /= 1e9;
x *= COBBLESTONE_FACTOR;
x /= SECONDS_PER_DAY;
x *= scale;
host.claimed_credit_per_cpu_sec = x;
if (config.granted_credit_weight) {
modify_credit_rating(host);
}
}
// modify host struct based on request.
// Copy all fields that are determined by the client.
//
static int modify_host_struct(HOST& host) {
host.timezone = g_request->host.timezone;
strncpy(host.domain_name, g_request->host.domain_name, sizeof(host.domain_name));
g_request->coprocs.summary_string(host.serialnum, sizeof(host.serialnum));
if (strcmp(host.last_ip_addr, g_request->host.last_ip_addr)) {
strncpy(host.last_ip_addr, g_request->host.last_ip_addr, sizeof(host.last_ip_addr));
} else {
host.nsame_ip_addr++;
}
host.on_frac = g_request->host.on_frac;
host.connected_frac = g_request->host.connected_frac;
host.active_frac = g_request->host.active_frac;
host.duration_correction_factor = g_request->host.duration_correction_factor;
host.p_ncpus = g_request->host.p_ncpus;
strncpy(host.p_vendor, g_request->host.p_vendor, sizeof(host.p_vendor));
// unlikely this will change
strncpy(host.p_model, g_request->host.p_model, sizeof(host.p_model));
host.p_fpops = g_request->host.p_fpops;
host.p_iops = g_request->host.p_iops;
host.p_membw = g_request->host.p_membw;
strncpy(host.os_name, g_request->host.os_name, sizeof(host.os_name));
strncpy(host.os_version, g_request->host.os_version, sizeof(host.os_version));
host.m_nbytes = g_request->host.m_nbytes;
host.m_cache = g_request->host.m_cache;
host.m_swap = g_request->host.m_swap;
host.d_total = g_request->host.d_total;
host.d_free = g_request->host.d_free;
host.d_boinc_used_total = g_request->host.d_boinc_used_total;
host.d_boinc_used_project = g_request->host.d_boinc_used_project;
host.n_bwup = g_request->host.n_bwup;
host.n_bwdown = g_request->host.n_bwdown;
if (strlen(g_request->host.host_cpid)) {
strcpy(host.host_cpid, g_request->host.host_cpid);
}
host.fix_nans();
compute_credit_rating(host);
g_request->host.claimed_credit_per_cpu_sec = host.claimed_credit_per_cpu_sec;
return 0;
}
// update the DB record to the values in "xhost"
// "initial_host" stores the current DB values;
// update only those fields that have changed
//
static int update_host_record(HOST& initial_host, HOST& xhost, USER& user) {
DB_HOST host;
int retval;
char buf[1024];
host = xhost;
// hash the CPID reported by the host with the user's email address.
// This prevents one user from spoofing another one's host.
//
if (strlen(host.host_cpid)) {
sprintf(buf, "%s%s", host.host_cpid, user.email_addr);
md5_block((const unsigned char*)buf, strlen(buf), host.host_cpid);
}
char* p = getenv("REMOTE_ADDR");
if (p) {
strlcpy(host.external_ip_addr, p, sizeof(host.external_ip_addr));
}
retval = host.update_diff(initial_host);
if (retval) {
log_messages.printf(MSG_CRITICAL, "host.update() failed: %d\n", retval);
}
return 0;
}
inline const char* reason_str(int n) {
switch (n) {
case ABORT_REASON_NOT_FOUND: return "result not in request";
case ABORT_REASON_WU_CANCELLED: return "WU cancelled";
case ABORT_REASON_ASSIMILATED: return "WU assimilated";
case ABORT_REASON_TIMED_OUT: return "result timed out";
}
return "Unknown";
}
// Figure out which of the results the host currently has
// should be aborted outright, or aborted if not started yet
//
int send_result_abort() {
int aborts_sent = 0;
int retval = 0;
DB_IN_PROGRESS_RESULT result;
std::string result_names;
unsigned int i;
if (g_request->other_results.size() == 0) {
return 0;
}
// build list of result names
//
for (i=0; i<g_request->other_results.size(); i++) {
OTHER_RESULT& orp=g_request->other_results[i];
orp.abort = true;
orp.abort_if_not_started = false;
orp.reason = ABORT_REASON_NOT_FOUND;
if (i > 0) result_names.append(", ");
result_names.append("'");
result_names.append(orp.name);
result_names.append("'");
}
// look up selected fields from the results and their WUs,
// and decide if they should be aborted
//
while (!(retval = result.enumerate(g_reply->host.id, result_names.c_str()))) {
for (i=0; i<g_request->other_results.size(); i++) {
OTHER_RESULT& orp = g_request->other_results[i];
if (!strcmp(orp.name, result.result_name)) {
if (result.error_mask&WU_ERROR_CANCELLED ) {
// if the WU has been canceled, abort the result
//
orp.abort = true;
orp.abort_if_not_started = false;
orp.reason = ABORT_REASON_WU_CANCELLED;
} else if ( result.assimilate_state == ASSIMILATE_DONE ) {
// if the WU has been assimilated, abort if not started
//
orp.abort = false;
orp.abort_if_not_started = true;
orp.reason = ABORT_REASON_ASSIMILATED;
} else if (result.server_state == RESULT_SERVER_STATE_OVER
&& result.outcome == RESULT_OUTCOME_NO_REPLY
) {
// if timed out, abort if not started
//
orp.abort = false;
orp.abort_if_not_started = true;
orp.reason = ABORT_REASON_TIMED_OUT;
} else {
// all is good with the result - let it process
orp.abort = false;
orp.abort_if_not_started = false;
}
break;
}
}
}
// If enumeration returned an error, don't send any aborts
//
if (retval && (retval != ERR_DB_NOT_FOUND)) {
return retval;
}
// loop through the results and send the appropriate message (if any)
//
for (i=0; i<g_request->other_results.size(); i++) {
OTHER_RESULT& orp = g_request->other_results[i];
if (orp.abort) {
g_reply->result_aborts.push_back(orp.name);
log_messages.printf(MSG_NORMAL,
"[HOST#%d]: Send result_abort for result %s; reason: %s\n",
g_reply->host.id, orp.name, reason_str(orp.reason)
);
// send user message
char buf[256];
sprintf(buf, "Result %s is no longer usable", orp.name);
g_reply->insert_message(USER_MESSAGE(buf, "high"));
} else if (orp.abort_if_not_started) {
g_reply->result_abort_if_not_starteds.push_back(orp.name);
log_messages.printf(MSG_NORMAL,
"[HOST#%d]: Send result_abort_if_unstarted for result %s; reason %d\n",
g_reply->host.id, orp.name, orp.reason
);
}
}
return aborts_sent;
}
// 1) Decide which global prefs to use for sched decisions: either
// - <working_global_prefs> from request msg
// - <global_prefs> from request message
// - prefs from user DB record
// and parse them into g_request->global_prefs.
// 2) update prefs in user record if needed
// 2) send global prefs in reply msg if needed
//
int handle_global_prefs() {
char buf[BLOB_SIZE];
g_reply->send_global_prefs = false;
bool have_working_prefs = (strlen(g_request->working_global_prefs_xml)>0);
bool have_master_prefs = (strlen(g_request->global_prefs_xml)>0);
bool have_db_prefs = (strlen(g_reply->user.global_prefs)>0);
bool same_account = !strcmp(
g_request->global_prefs_source_email_hash, g_reply->email_hash
);
double master_mod_time=0, db_mod_time=0;
if (have_master_prefs) {
parse_double(g_request->global_prefs_xml, "<mod_time>", master_mod_time);
if (master_mod_time > dtime()) master_mod_time = dtime();
}
if (have_db_prefs) {
parse_double(g_reply->user.global_prefs, "<mod_time>", db_mod_time);
if (db_mod_time > dtime()) db_mod_time = dtime();
}
if (config.debug_prefs) {
log_messages.printf(MSG_DEBUG,
"have_master:%d have_working: %d have_db: %d\n",
have_master_prefs, have_working_prefs, have_db_prefs
);
}
// decide which prefs to use for sched decisions,
// and parse them into g_request->global_prefs
//
if (have_working_prefs) {
g_request->global_prefs.parse(g_request->working_global_prefs_xml, "");
if (config.debug_prefs) {
log_messages.printf(MSG_DEBUG, "using working prefs\n");
}
} else {
if (have_master_prefs) {
if (have_db_prefs && db_mod_time > master_mod_time) {
g_request->global_prefs.parse(g_reply->user.global_prefs, g_reply->host.venue);
if (config.debug_prefs) {
log_messages.printf(MSG_DEBUG, "using db prefs - more recent\n");
}
} else {
g_request->global_prefs.parse(g_request->global_prefs_xml, g_reply->host.venue);
if (config.debug_prefs) {
log_messages.printf(MSG_DEBUG, "using master prefs\n");
}
}
} else {
if (have_db_prefs) {
g_request->global_prefs.parse(g_reply->user.global_prefs, g_reply->host.venue);
if (config.debug_prefs) {
log_messages.printf(MSG_DEBUG, "using db prefs\n");
}
} else {
g_request->global_prefs.defaults();
if (config.debug_prefs) {
log_messages.printf(MSG_DEBUG, "using default prefs\n");
}
}
}
}
// decide whether to update DB
//
if (have_master_prefs) {
bool update_user_record = false;
if (have_db_prefs) {
if (master_mod_time > db_mod_time && same_account) {
update_user_record = true;
}
} else {
if (same_account) update_user_record = true;
}
if (update_user_record) {
log_messages.printf(MSG_DEBUG, "updating db prefs\n");
strcpy(g_reply->user.global_prefs, g_request->global_prefs_xml);
DB_USER user;
user.id = g_reply->user.id;
escape_string(g_request->global_prefs_xml, sizeof(g_request->global_prefs_xml));
sprintf(buf, "global_prefs='%s'", g_request->global_prefs_xml);
unescape_string(g_request->global_prefs_xml, sizeof(g_request->global_prefs_xml));
int retval = user.update_field(buf);
if (retval) {
log_messages.printf(MSG_CRITICAL,
"user.update_field() failed: %d\n", retval
);
}
}
}
// decide whether to send DB prefs in reply msg
//
if (config.debug_prefs) {
log_messages.printf(MSG_DEBUG,
"have db %d; dbmod %f; global mod %f\n",
have_db_prefs, db_mod_time, g_request->global_prefs.mod_time
);
}
if (have_db_prefs && db_mod_time > master_mod_time) {
if (config.debug_prefs) {
log_messages.printf(MSG_DEBUG, "sending db prefs in reply\n");
}
g_reply->send_global_prefs = true;
}
return 0;
}
// if the client has an old code sign public key,
// send it the new one, with a signature based on the old one.
// If they don't have a code sign key, send them one
//
bool send_code_sign_key(char* code_sign_key) {
char* oldkey, *signature;
int i, retval;
char path[256];
if (strlen(g_request->code_sign_key)) {
if (strcmp(g_request->code_sign_key, code_sign_key)) {
log_messages.printf(MSG_NORMAL, "received old code sign key\n");
// look for a signature file
//
for (i=0; ; i++) {
sprintf(path, "%s/old_key_%d", config.key_dir, i);
retval = read_file_malloc(path, oldkey);
if (retval) {
g_reply->insert_message(
USER_MESSAGE(
"You may have an outdated code verification key. "
"This may prevent you from accepting new executables. "
"If the problem persists, detach/attach the project. ",
"high"
)
);
return false;
}
if (!strcmp(oldkey, g_request->code_sign_key)) {
sprintf(path, "%s/signature_%d", config.key_dir, i);
retval = read_file_malloc(path, signature);
if (retval) {
g_reply->insert_message(
USER_MESSAGE(
"You may have an outdated code verification key. "
"This may prevent you from accepting new executables. "
"If the problem persists, detach/attach the project. ",
"high"
)
);
} else {
safe_strcpy(g_reply->code_sign_key, code_sign_key);
safe_strcpy(g_reply->code_sign_key_signature, signature);
free(signature);
}
}
free(oldkey);
return false;
}
}
} else {
safe_strcpy(g_reply->code_sign_key, code_sign_key);
}
return true;
}
// This routine examines the <min_core_client_version_announced> value
// from config.xml. If set, and the core client version is less than
// this version, send a warning to users to upgrade before deadline
// given in <min_core_client_upgrade_deadline> in Unix time(2) format
// expires.
//
void warn_user_if_core_client_upgrade_scheduled() {
if (g_request->core_client_version < config.min_core_client_version_announced) {
// time remaining in hours, before upgrade required
int remaining = config.min_core_client_upgrade_deadline-time(0);
remaining /= 3600;
if (0 < remaining) {
char msg[512];
int days = remaining / 24;
int hours = remaining % 24;
sprintf(msg,
"Starting in %d days and %d hours, project will require a minimum "
"BOINC core client version of %d.%d.0. You are currently using "
"version %d.%d.%d; please upgrade before this time.",
days, hours,
config.min_core_client_version_announced / 100,
config.min_core_client_version_announced % 100,
g_request->core_client_major_version,
g_request->core_client_minor_version,
g_request->core_client_release
);
// make this low priority until three days are left. Then
// bump to high.
//
if (days<3) {
g_reply->insert_message(USER_MESSAGE(msg, "high"));
} else {
g_reply->insert_message(USER_MESSAGE(msg, "low"));
}
log_messages.printf(MSG_DEBUG,
"Sending warning: upgrade client %d.%d.%d within %d days %d hours\n",
g_request->core_client_major_version,
g_request->core_client_minor_version,
g_request->core_client_release,
days, hours
);
}
}
return;
}
bool unacceptable_os() {
unsigned int i;
char buf[1024];
for (i=0; i<config.ban_os->size(); i++) {
regex_t& re = (*config.ban_os)[i];
strcpy(buf, g_request->host.os_name);
strcat(buf, "\t");
strcat(buf, g_request->host.os_version);
if (!regexec(&re, buf, 0, NULL, 0)) {
log_messages.printf(MSG_NORMAL,
"Unacceptable OS %s %s\n",
g_request->host.os_name, g_request->host.os_version
);
sprintf(buf, "This project doesn't support OS type %s %s",
g_request->host.os_name, g_request->host.os_version
);
g_reply->insert_message(USER_MESSAGE(buf, "low"));
g_reply->set_delay(DELAY_UNACCEPTABLE_OS);
return true;
}
}
return false;
}
bool unacceptable_cpu() {
unsigned int i;
char buf[1024];
for (i=0; i<config.ban_cpu->size(); i++) {
regex_t& re = (*config.ban_cpu)[i];
strcpy(buf, g_request->host.p_vendor);
strcat(buf, "\t");
strcat(buf, g_request->host.p_model);
if (!regexec(&re, buf, 0, NULL, 0)) {
log_messages.printf(MSG_NORMAL,
"Unacceptable CPU %s %s\n",
g_request->host.p_vendor, g_request->host.p_model
);
sprintf(buf, "This project doesn't support CPU type %s %s",
g_request->host.p_vendor, g_request->host.p_model
);
g_reply->insert_message(USER_MESSAGE(buf, "low"));
g_reply->set_delay(DELAY_UNACCEPTABLE_OS);
return true;
}
}
return false;
}
bool wrong_core_client_version() {
char msg[256];
bool wrong_version = false;
if (config.min_core_client_version) {
int major = config.min_core_client_version/100;
int minor = config.min_core_client_version % 100;
if (g_request->core_client_major_version < major ||
((g_request->core_client_major_version == major) && (g_request->core_client_minor_version < minor))) {
wrong_version = true;
sprintf(msg,
"Need version %d.%d or higher of the BOINC core client. You have %d.%d.",
major, minor,
g_request->core_client_major_version, g_request->core_client_minor_version
);
log_messages.printf(MSG_NORMAL,
"[HOST#%d] [auth %s] Wrong minor version from user: wanted %d, got %d\n",
g_request->hostid, g_request->authenticator,
minor, g_request->core_client_minor_version
);
}
}
if (wrong_version) {
g_reply->insert_message(USER_MESSAGE(msg, "low"));
g_reply->set_delay(DELAY_BAD_CLIENT_VERSION);
return true;
}
return false;
}
inline static const char* get_remote_addr() {
const char * r = getenv("REMOTE_ADDR");
return r ? r : "?.?.?.?";
}
void handle_msgs_from_host() {
unsigned int i;
DB_MSG_FROM_HOST mfh;
int retval;
for (i=0; i<g_request->msgs_from_host.size(); i++) {
g_reply->send_msg_ack = true;
MSG_FROM_HOST_DESC& md = g_request->msgs_from_host[i];
mfh.clear();
mfh.create_time = time(0);
safe_strcpy(mfh.variety, md.variety);
mfh.hostid = g_reply->host.id;
mfh.handled = false;
safe_strcpy(mfh.xml, md.msg_text.c_str());
log_messages.printf(MSG_NORMAL,
"got msg from host; variety %s \n",
mfh.variety
);
retval = mfh.insert();
if (retval) {
log_messages.printf(MSG_CRITICAL,
"[HOST#%d] message insert failed: %d\n",
g_reply->host.id, retval
);
g_reply->send_msg_ack = false;
// may as well return; if one insert failed, others will too
//
return;
}
}
}
void handle_msgs_to_host() {
DB_MSG_TO_HOST mth;
char buf[256];
sprintf(buf, "where hostid = %d and handled = %d", g_reply->host.id, 0);
while (!mth.enumerate(buf)) {
g_reply->msgs_to_host.push_back(mth);
mth.handled = true;
mth.update();
}
}
static void log_request() {
log_messages.printf(MSG_NORMAL,
"Request: [USER#%d] [HOST#%d] [IP %s] client %d.%d.%d\n",
g_reply->user.id, g_reply->host.id, get_remote_addr(),
g_request->core_client_major_version,
g_request->core_client_minor_version,
g_request->core_client_release
);
if (config.debug_request_details) {
log_messages.printf(MSG_DEBUG,
"Request details: auth %s, RPC seqno %d, platform %s\n",
g_request->authenticator,
g_request->rpc_seqno,
g_request->platform.name
);
}
log_messages.set_indent_level(2);
}
bool bad_install_type() {
if (config.no_vista_sandbox) {
if (!strcmp(g_request->host.os_name, "Microsoft Windows Vista")) {
if (g_request->sandbox == 1) {
log_messages.printf(MSG_NORMAL,
"Vista secure install - not sending work\n"
);
g_reply->insert_message(
USER_MESSAGE(
"Unable to send work to Vista with BOINC installed in protected mode",
"high"
)
);
g_reply->insert_message(
USER_MESSAGE(
"Please reinstall BOINC and uncheck 'Protected application execution'",
"high"
)
);
}
}
}
return false;
}
static inline bool requesting_work() {
if (g_request->work_req_seconds > 0) return true;
if (g_request->cpu_req_secs > 0) return true;
if (g_request->coproc_cuda && g_request->coproc_cuda->req_secs) return true;
return false;
}
void process_request(char* code_sign_key) {
PLATFORM* platform;
int retval;
double last_rpc_time;
struct tm *rpc_time_tm;
int last_rpc_dayofyear;
int current_rpc_dayofyear;
bool ok_to_send_work = true;
bool have_no_work;
char buf[256];
HOST initial_host;
unsigned int i;
time_t t;
memset(&g_reply->wreq, 0, sizeof(g_reply->wreq));
// if different major version of BOINC, just send a message
//
if (wrong_core_client_version()
|| unacceptable_os()
|| unacceptable_cpu()
) {
ok_to_send_work = false;
// if no results, return without accessing DB
//
if (g_request->results.size() == 0) {
return;
}
} else {
warn_user_if_core_client_upgrade_scheduled();
}
if (requesting_work()) {
if (config.locality_scheduling || config.locality_scheduler_fraction || config.enable_assignment) {
have_no_work = false;
} else {
lock_sema();
have_no_work = ssp->no_work(g_pid);
if (have_no_work) {
g_wreq->no_jobs_available = true;
log_messages.printf(MSG_NORMAL, "No jobs in shmem\n");
}
unlock_sema();
}
}
// if there's no work, and client isn't returning results,
// this isn't an initial RPC,
// and client is requesting work, return without accessing DB
//
if (
config.nowork_skip
&& requesting_work()
&& have_no_work
&& (g_request->results.size() == 0)
&& (g_request->hostid != 0)
) {
g_reply->insert_message(USER_MESSAGE("No work available", "low"));
g_reply->set_delay(DELAY_NO_WORK_SKIP);
if (!config.msg_to_host) {
log_messages.printf(MSG_NORMAL, "No work - skipping DB access\n");
return;
}
}
// FROM HERE ON DON'T RETURN; goto leave instead
// because we've tagged an entry in the work array with our process ID
// now open the database
//
retval = open_database();
if (retval) {
send_error_message("Server can't open database", 3600);
g_reply->project_is_down = true;
goto leave;
}
retval = authenticate_user();
if (retval) goto leave;
if (g_reply->user.id == 0) {
log_messages.printf(MSG_CRITICAL, "No user ID!\n");
}
initial_host = g_reply->host;
g_reply->host.rpc_seqno = g_request->rpc_seqno;
g_reply->nucleus_only = false;
log_request();
// is host blacklisted?
//
if (g_reply->host.max_results_day == -1) {
send_error_message("Not accepting requests from this host", 86400);
goto leave;
}
if (strlen(config.sched_lockfile_dir)) {
int pid_with_lock = lock_sched();
if (pid_with_lock > 0) {
log_messages.printf(MSG_CRITICAL,
"Another scheduler instance [PID=%d] is running for this host\n",
pid_with_lock
);
} else if (pid_with_lock) {
log_messages.printf(MSG_CRITICAL,
"Error acquiring lock for [HOST#%d]\n", g_reply->host.id
);
}
if (pid_with_lock) {
send_error_message(
"Another scheduler instance is running for this host", 60
);
goto leave;
}
}
last_rpc_time = g_reply->host.rpc_time;
t = g_reply->host.rpc_time;
rpc_time_tm = localtime(&t);
last_rpc_dayofyear = rpc_time_tm->tm_yday;
t = time(0);
g_reply->host.rpc_time = t;
rpc_time_tm = localtime(&t);
current_rpc_dayofyear = rpc_time_tm->tm_yday;
if (config.daily_result_quota) {
if (g_reply->host.max_results_day == 0 || g_reply->host.max_results_day > config.daily_result_quota) {
g_reply->host.max_results_day = config.daily_result_quota;
log_messages.printf(MSG_DEBUG,
"[HOST#%d] Initializing max_results_day to %d\n",
g_reply->host.id, config.daily_result_quota
);
}
}
if (last_rpc_dayofyear != current_rpc_dayofyear) {
log_messages.printf(MSG_DEBUG,
"[HOST#%d] Resetting nresults_today\n", g_reply->host.id
);
g_reply->host.nresults_today = 0;
}
retval = modify_host_struct(g_reply->host);
// write time stats to disk if present
//
if (g_request->have_time_stats_log) {
write_time_stats_log();
}
// look up the client's platform(s) in the DB
//
platform = ssp->lookup_platform(g_request->platform.name);
if (platform) g_request->platforms.list.push_back(platform);
for (i=0; i<g_request->alt_platforms.size(); i++) {
platform = ssp->lookup_platform(g_request->alt_platforms[i].name);
if (platform) g_request->platforms.list.push_back(platform);
}
if (g_request->platforms.list.size() == 0) {
sprintf(buf, "platform '%s' not found", g_request->platform.name);
g_reply->insert_message(USER_MESSAGE(buf, "low"));
log_messages.printf(MSG_CRITICAL,
"[HOST#%d] platform '%s' not found\n",
g_reply->host.id, g_request->platform.name
);
g_reply->set_delay(DELAY_PLATFORM_UNSUPPORTED);
goto leave;
}
handle_global_prefs();
handle_results();
// Do this before resending lost jobs
//
if (bad_install_type()) {
ok_to_send_work = false;
}
send_work_setup();
g_wreq->njobs_on_host = g_request->other_results.size();
for (i=0; i<g_request->other_results.size(); i++) {
OTHER_RESULT& r = g_request->other_results[i];
if (r.have_plan_class) {
if (app_plan_uses_gpu(r.plan_class)) {
g_wreq->njobs_on_host_gpu++;
} else {
g_wreq->njobs_on_host_cpu++;
}
}
}
if (g_request->have_other_results_list) {
if (config.resend_lost_results && ok_to_send_work) {
if (resend_lost_work()) {
ok_to_send_work = false;
}
}
if (config.send_result_abort) {
send_result_abort();
}
}
if (requesting_work()) {
if (!send_code_sign_key(code_sign_key)) {
ok_to_send_work = false;
}
// if last RPC was within config.min_sendwork_interval, don't send work
//
if (!have_no_work && ok_to_send_work) {
if (config.min_sendwork_interval) {
double diff = dtime() - last_rpc_time;
if (diff < config.min_sendwork_interval) {
ok_to_send_work = false;
log_messages.printf(MSG_NORMAL,
"Not sending work - last request too recent: %f\n", diff
);
sprintf(buf,
"Not sending work - last request too recent: %d sec", (int)diff
);
g_reply->insert_message(USER_MESSAGE(buf, "low"));
// the 1.01 is in case client's clock
// is slightly faster than ours
//
g_reply->set_delay(1.01*config.min_sendwork_interval);
}
}
if (ok_to_send_work) {
send_work();
}
}
if (g_wreq->no_jobs_available) {
g_reply->insert_message(
USER_MESSAGE(
"(Project has no jobs available)",
"high"
)
);
}
}
handle_msgs_from_host();
if (config.msg_to_host) {
handle_msgs_to_host();
}
update_host_record(initial_host, g_reply->host, g_reply->user);
leave:
if (!have_no_work) {
ssp->restore_work(g_pid);
}
}
static void log_incomplete_request() {
// BOINC scheduler requests use method POST.
// So method GET means that someone is trying a browser.
//
char *rm=getenv("REQUEST_METHOD");
bool used_get = false;
if (rm && !strcmp(rm, "GET")) {
used_get = true;
}
log_messages.printf(MSG_NORMAL,
"Incomplete request received %sfrom IP %s, auth %s, platform %s, version %d.%d.%d\n",
used_get?"(used GET method - probably a browser) ":"",
get_remote_addr(), g_request->authenticator, g_request->platform.name,
g_request->core_client_major_version, g_request->core_client_minor_version,
g_request->core_client_release
);
}
static void log_user_messages() {
for (unsigned int i=0; i<g_reply->messages.size(); i++) {
USER_MESSAGE um = g_reply->messages[i];
log_messages.printf(MSG_DEBUG,
"[HOST#%d] MSG(%4s) %s\n",
g_reply->host.id, um.priority.c_str(), um.message.c_str()
);
}
}
void handle_request(FILE* fin, FILE* fout, char* code_sign_key) {
SCHEDULER_REQUEST sreq;
SCHEDULER_REPLY sreply;
char buf[1024];
double start_time = dtime();
g_request = &sreq;
g_reply = &sreply;
g_wreq = &sreply.wreq;
memset(&sreq, 0, sizeof(sreq));
sreply.nucleus_only = true;
log_messages.set_indent_level(1);
const char* p = sreq.parse(fin);
if (!p){
process_request(code_sign_key);
} else {
sprintf(buf, "Error in request message: %s", p);
log_incomplete_request();
sreply.insert_message(USER_MESSAGE(buf, "low"));
}
if ((config.locality_scheduling || config.locality_scheduler_fraction) && !sreply.nucleus_only) {
send_file_deletes();
}
if (config.debug_user_messages) {
log_user_messages();
}
sreply.write(fout, sreq);
log_messages.printf(MSG_NORMAL,
"Scheduler ran %.3f seconds\n", dtime()-start_time
);
if (strlen(config.sched_lockfile_dir)) {
unlock_sched();
}
}
const char *BOINC_RCSID_2ac231f9de = "$Id$";
Reference in New Issue View Git Blame Copy Permalink