W przypadku Axis-a można to rozwiązać tak:
System.setProperty("tcp.window.size", "64000");
AxisProperties.setProperty("axis.socketFactory", "CustomSocketFactory");
Czasy WebService'u dla rozmiaru okna:
1460 (RWIN=MSS :) => 21860 ms27109 adinstan 469M 323M cpu1 30 0 209:37:17 35% bwengine/64Proces skacze po procesorach. Można go dowiązać narzędziami psrset+pbind, co powinno zmniejszyć opóźnienia (ważna rzecz dla np. aplikacji RT).
27109 adinstan 469M 323M cpu17 20 0 209:37:45 35% bwengine/64
27109 adinstan 469M 323M cpu16 20 0 209:38:14 35% bwengine/64
27109 adinstan 469M 323M cpu17 20 0 209:38:56 35% bwengine/64
27109 adinstan 469M 323M cpu18 20 0 209:39:10 35% bwengine/64
27109 adinstan 469M 323M cpu0 10 0 209:39:24 35% bwengine/64
27109 adinstan 469M 323M cpu2 10 0 209:39:38 35% bwengine/64
27109 adinstan 469M 323M cpu19 10 0 209:40:07 35% bwengine/64
public GetSetAccessor(ClassNo i gdzie to interoperability out of box?...
baseUrl = Clazz.class.getResource(".");
url = new URL(baseURL, "clazz.wsdl");należy zamienić na:url = Clazz.class.getResource("clazz.wsdl");
new weblogic.security.internal.encryption.ClearOrEncryptedService(
SerializedSystemIni.getEncryptionService(
"/home/weblogic/bea/domains/wls0"
)
).decrypt("{3DES}1234abcd==");
HOSTNAME=`hostname`
USER=`whoami`
SMTP="10.10.1.25"
function sendmail {
(
sleep 3
echo "HELO $HOSTNAME"
sleep 1
echo "mail from: monitoring@email.com"
sleep 1
echo "rcpt to: smsgw@email.com"
sleep 1
echo "data"
sleep 1
echo "From: monitoring.$USER@email.com"
sleep 1
echo "To: smsgw@email.com"
sleep 1
echo "Subject: $1"
sleep 1
echo ""
sleep 1
echo "$2"
sleep 1
echo "."
sleep 1
echo "quit"
sleep 1
) | telnet $SMTP 25
}
sendmail "$1" "$2"
# ./sendsms "+48602201122" "Brak dzialajacego procesu A dla systemu B"
WP_URL="http://10.10.1.20:8080/app/AppServlet"
COOKIE="cookie.dat"
WGET_CMD="wget --timeout=15 --keep-session-cookies --save-cookies=$COOKIE --quiet $WP_URL -O -"
> $COOKIE
RUNNING=$(ps uax | grep -c nazwa_skryptu.sh)
[[ $RUNNING -gt 3 ]] && echo "Poprzednie wywolanie skryptu wisi" && exit 1
RES=$($WGET_CMD --post-data="user=jan.kowalski&password=lubieTpsa&action=login" | grep -c "Zalogowany: ")
echo "Test logowania. Wynik: $RES"
RES=$($WGET_CMD --load-cookies=$COOKIE --post-data="action=create-process&action_params=...&state=" | grep -c "Nowy proces")
echo "Zaladowanie procesu. Wynik: $RES"
RES=$($WGET_CMD --load-cookies=$COOKIE --post-data="action=next&action_params=..." | grep -c "Dane zapisano")
echo "Wynik procesu poprawny?: $RES"
RES=$($WGET_CMD --load-cookies=$COOKIE --post-data="action=save-and-close&action_params=..." | grep -c "Zakonczono proces")
echo "Zamkniecie procesu. Wynik: $RES"
RES=$($WGET_CMD --load-cookies=$COOKIE --post-data="action=logout" | grep -c "Wylogowano uzytkownika")
echo "Wylogowanie. Wynik: $RES"
#!/bin/sh
RATE_LIMIT="100"
HOSTS_DENY=".htaccess"
LOG="/home/apache_wp/apache_wp/logs/access_log_443"
DT=$(date +%d/%b/%Y:%k:%M)
PERIODIC_LOG="periodic.log"
cat $LOG | grep "$DT" | grep "SomeServletUrl" | cut -d" " -f1 > $PERIODIC_LOG
echo "### Generated at $DT ###">> $HOSTS_DENY
for ip in $(cat $PERIODIC_LOG | sort -u); do
cnt=$(cat $PERIODIC_LOG | grep -c "$ip");
[[ $cnt -gt $RATE_LIMIT ]] || continue;
echo >> $HOSTS_DENY;
echo "# $ip generated $cnt requests above rate limit $RATE_LIMIT" >> $HOSTS_DENY;
echo "Deny from $ip" >> $HOSTS_DENY;
done;
#define _GNU_SOURCE
#include <stdio.h>
#include <stdlib.h>
#include <unistd.h>
#include <sched.h>
#include <errno.h>
#include <string.h>
int get_num_cpus() {
int cnt = 0;
char *line = NULL;
size_t len = 0;
FILE *fp = fopen("/proc/stat", "r");
if (fp!=NULL) {
while (getline(&line, &len, fp)!=-1) {
if (line!=NULL && strstr(line, "cpu")!=NULL)
cnt++;
}
if (line)
free(line);
fclose(fp);
cnt--;
}
return cnt;
}
int main(int argc, char** argv) {
static cpu_set_t mask;
int user_mask, i=0, max, m=0;
if (argc<3) {
fprintf(stdout, "Usage: %s cpu_mask(int) command(full_path)\n",argv[0]);
exit(0);
}
user_mask = atoi(argv[1]);
max = 1 << (get_num_cpus()-1);
CPU_ZERO(&mask);
for (i=0; i<max; i++) {
if ( (user_mask & (1 << i)) == (1 << i)) {
CPU_SET(i, &mask);
m |= (1<<i);
}
}
if (sched_setaffinity(0, sizeof(mask), &mask)==0) {
fprintf(stdout, "Running %s with mask %d, cpus=%d\n", argv[2], m, max);
return execv(argv[2], argv+2);
}
else {
fprintf(stderr, "Cannot set cpu mask: %s. Effective mask was %d\n", strerror(errno), m);
exit(1);
}
}
[tomcat@swa158 ~]$ jrcmd 0 print_memusageTomcaty są skonfigurowane z -Xmx384M a pamięć wirtualna (VSZ) dla procesu po dwóch tygodniach wynosi 1192MB, RSS 900MB. Czyli o 500-800MB za dużo. Sama diagnostyka JRockita pokazuje unaccounted for memory 91.3%. Ta sama aplikacja chodząca bez Wily’ego zachowuje się poprawnie. Średnio na dzień wycieka około 50MB pamięci.
18130:
[JRockit] memtrace is collecting data...
[JRockit] *** 0th memory utilization report
(all numbers are in kbytes)
Total mapped ;;;;;;;1192600
; Total in-use ;;;;;;1182252
;; executable ;;;;; 12260
;;; java code ;;;; 7680; 62.6%
;;;; used ;;; 5763; 75.0%
;; shared modules (exec+ro+rw) ;;;;; 4540
;; guards ;;;;; 1056
;; readonly ;;;;; 68
;; rw-memory ;;;;;1177604
;;; Java-heap ;;;; 393216; 33.4%
;;; Stacks ;;;; 22871; 1.9%
;;; Native-memory ;;;; 761516; 64.7%
;;;; java-heap-overhead ;;; 12342
;;;; codegen memory ;;; 904
;;;; classes ;;; 29696; 3.9%
;;;;; method bytecode ;; 2927
;;;;; method structs ;; 3278 (#69939)
;;;;; constantpool ;; 11117
;;;;; classblock ;; 1488
;;;;; class ;; 3060 (#8698)
;;;;; other classdata ;; 4942
;;;;; overhead ;; 280
;;;; threads ;;; 107; 0.0%
;;;; malloc:ed memory ;;; 24265; 3.2%
;;;;; codeinfo ;; 1047
;;;;; codeinfotrees ;; 438
;;;;; exceptiontables ;; 591
;;;;; metainfo/livemaptable ;; 9324
;;;;; codeblock structs ;; 3
;;;;; constants ;; 6
;;;;; livemap global tables ;; 2091
;;;;; callprof cache ;; 0
;;;;; paraminfo ;; 88 (#1284)
;;;;; strings ;; 4991 (#92449)
;;;;; strings(jstring) ;; 0
;;;;; typegraph ;; 1593
;;;;; interface implementor list ;; 48
;;;;; thread contexts ;; 81
;;;;; jar/zip memory ;; 5451
;;;;; native handle memory ;; 39
;;;; unaccounted for memory ;;; 694309; 91.2%;28.61
---------------------!!!
16612:
[JRockit] memtrace is collecting data...
[JRockit] *** 0th memory utilization report
(all numbers are in kbytes)
Total mapped ;;;;;;;1192932
; Total in-use ;;;;;;1181960
;; executable ;;;;; 12004
;;; java code ;;;; 7424; 61.8%
;;;; used ;;; 5796; 78.1%
;; shared modules (exec+ro+rw) ;;;;; 4540
;; guards ;;;;; 1068
;; readonly ;;;;; 68
;; rw-memory ;;;;;1177312
;;; Java-heap ;;;; 393216; 33.4%
;;; Stacks ;;;; 23003; 2.0%
;;; Native-memory ;;;; 761092; 64.6%
;;;; java-heap-overhead ;;; 12342
;;;; codegen memory ;;; 592
;;;; classes ;;; 28928; 3.8%
;;;;; method bytecode ;; 2911
;;;;; method structs ;; 3256 (#69475)
;;;;; constantpool ;; 10845
;;;;; classblock ;; 1441
;;;;; class ;; 2938 (#8350)
;;;;; other classdata ;; 4923
;;;;; overhead ;; 61
;;;; threads ;;; 109; 0.0%
;;;; malloc:ed memory ;;; 24479; 3.2%
;;;;; codeinfo ;; 1019
;;;;; codeinfotrees ;; 420
;;;;; exceptiontables ;; 585
;;;;; metainfo/livemaptable ;; 9489
;;;;; codeblock structs ;; 3
;;;;; constants ;; 7
;;;;; livemap global tables ;; 2183
;;;;; callprof cache ;; 0
;;;;; paraminfo ;; 89 (#1288)
;;;;; strings ;; 4997 (#92452)
;;;;; strings(jstring) ;; 0
;;;;; typegraph ;; 1527
;;;;; interface implementor list ;; 47
;;;;; thread contexts ;; 82
;;;;; jar/zip memory ;; 5451
;;;;; native handle memory ;; 41
;;;; unaccounted for memory ;;; 694751; 91.3%;28.38
---------------------!!!
[tomcat@swa158 ~]$
tomcat@swa175:~/tomcatDiag1/wily> find . -name *.so
./ext/libIntroscopeAIX4PSeries32Stats.so
./ext/libIntroscopeSolarisSparc64Stats.so
./ext/libIntroscopeRedHatStats.so
./ext/libIntroscopeHpuxParisc32Stats.so
./ext/libIntroscopeSolarisSparc32Stats.so
./ext/libIntroscopeSolarisAmd64Stats.so
./ext/libIntroscopeAIX52PSeries64Stats.so
./ext/libIntroscopeAIX53PSeries64Stats.so
./ext/libIntroscopeSolarisAmd32Stats.so
./ext/libIntroscopeHpuxItaniumStats.so
./ext/libIntroscopeAIX5PSeries32Stats.so
./ext/libIntroscopeAIX4PSeries64Stats.so
tomcat@swa175:~/tomcatDiag1/wily> nm ./ext/libIntroscopeRedHatStats.so | grep Java
000015e2 T Java_com_wily_introscope_agent_platform_linux_LinuxPlatformStatisticsBackEnd_getAggregateCPUUsage
0000158e T Java_com_wily_introscope_agent_platform_linux_LinuxPlatformStatisticsBackEnd_getParentProcessID
000011c6 T Java_com_wily_introscope_agent_platform_linux_LinuxPlatformStatisticsBackEnd_getProcessCPUUsage
0000153a T Java_com_wily_introscope_agent_platform_linux_LinuxPlatformStatisticsBackEnd_getProcessID
00001524 T Java_com_wily_introscope_agent_platform_linux_LinuxPlatformStatisticsBackEnd_getProtocolVersion
000017a0 T Java_com_wily_introscope_agent_platform_linux_LinuxPlatformStatisticsBackEnd_getThreadCpuTime
0000148a T Java_com_wily_introscope_agent_platform_linux_LinuxPlatformStatisticsBackEnd_getTotalClockTicks
0000152e T Java_com_wily_introscope_agent_platform_linux_LinuxPlatformStatisticsBackEnd_init
00001796 T Java_com_wily_introscope_agent_platform_linux_LinuxPlatformStatisticsBackEnd_isThreadCpuTimeEnabled
00001534 T Java_com_wily_introscope_agent_platform_linux_LinuxPlatformStatisticsBackEnd_term
private native long[] getAggregateCPUUsage(String s)
throws PlaformStatUnavailableException;
public class PH {
private static byte[] hash(String in, String key) throws Exception {
MessageDigest md = MessageDigest.getInstance("MD5");
md.update(key.getBytes());
return md.digest(in.getBytes());
}
private static String fillNumber(String s, int len) {
while (s.length() < len)
s = "0"+s;
return s;
}
private static String byte2String(byte[] b, int len) {
StringBuffer sb = new StringBuffer();
for (int i=0; i < Math.min(b.length, len); i++) {
int c = '0' + (b[i] & 0xff)%10;
sb.append((char)c);
}
return sb.toString();
}
public static void main(String[] args) throws Exception {
HashMap<String, String> map = new HashMap<String, String>();
int hc = 0;
for (int i=0; i < 100000; i++) {
String pesel = "721028" + fillNumber(i+"", 5);
String hash5 = byte2String( hash(pesel, "random123!"), 10);
if (map.get(hash5)!=null) {
System.err.println("Kolizja dla peselu "+pesel+" => "+hash5 + " => "+map.get(hash5));
hc++;
}
else
map.put(hash5, pesel);
}
System.err.println("Kolizji: "+hc);
}
}
CREATE OR REPLACE
PROCEDURE HASH_OSOBY AS
type ROWID_TAB is table of rowid index by binary_integer;
collisions rowid_tab;
collisions_cnt number := 0;
passes number := 0;
key varchar2(4000) := 'klucz1';
err varchar2(4000) := '';
BEGIN
dbms_output.disable;
for rec in (select rowid from monitor.os) loop
begin
update monitor.os set pesel = haszujpesel(pesel, key) where rowid=rec.rowid;
exception when others then
collisions_cnt := collisions_cnt+1;
collisions(collisions_cnt) := rec.rowid;
err := SQLERRM;
insert into log_h(T,LOG) values(sysdate(),err);
end;
end loop;
insert into log_h(T,LOG) values(sysdate(), 'Pass '||passes||', number of collisions: '||collisions_cnt);
while collisions.count > 0 loop
key := key||key;
for i in collisions.first..collisions.last loop
begin
update monitor.os set pesel = haszujpesel(pesel, key) where rowid=collisions(i);
collisions.delete(i);
exception when others then
err := SQLERRM;
insert into log_h(T,LOG) values(sysdate(),err);
end;
end loop;
passes := passes + 1;
collisions_cnt := collisions.count;
insert into log_h(T,LOG) values(sysdate(), 'Pass '||passes||', number of collisions: '||collisions_cnt);
end loop;
END HASH_OSOBY;
Node g = DOMHelper.getElementsByTag(p, "group-list").get(0);
rs = new DOMResultSet(g);
while (rs.next()) { // iteracja po węzłach poniżej group-list
// obiekt z wezła group
Group g = Group.fromResultSet(rs);
// w group wskazujemy węzły group-attrib
if (rs.findColumn("group-attrib")>0) {
while (rs.next()) {
GroupAttrib ga = GroupAttrib.fromResultSet(rs, g);
}
rs.refreshRow(); // kasujemy wskazanie
}
Debug.println("Wczytano "+rs.getString("."),this);
// "." oznacza biężący węzeł do zwrócenia jako string XML
}
Sterowniki to 185.18.08 beta. Czyli poprzedni krótki czas pracy na baterii to była wina NVidii.
Section "Device"
Identifier "Device0"
Driver "nvidia"
VendorName "NVIDIA Corporation"
Option "RegistryDwords" "PowerMizerLevel=0x1"
Option "Coolbits" "1"
EndSection
Da się zrobić, żeby wyglądało po staremu.
Ciekawostka: szybka zmiana rozmiaru osadzonego okna w Przeglądarce Zdarzeń powoduje wywalenie się mmc.exe. Trochę obciach jak na Release Candidate. Nie czuć optymalizacji GDI.
#include <windows.h>
#include <process.h>
int WINAPI WinMain (HINSTANCE hThisInstance,
HINSTANCE hPrevInstance,
LPSTR lpszArgument,
int nFunsterStil)
{
char* argv[32];
for (int i=0; i < 32; i++)
argv[i] = 0;
int argp = 2;
char *args = lpszArgument;
while(*args!=0) {
if (*args==' ') {
argv[argp++] = ++args;
args[-1] = 0;
}
}
argv[0] = "AppToFilterArgs.exe";
argv[1] = args;
return execv("AppToFilterArgs", argv);
}
import sun.misc.Unsafe;
public final class EvilCode {
public final static Unsafe getUnsafe () {
try {
Field field = Unsafe.class.getDeclaredField("theUnsafe");
field.setAccessible(true);
return (Unsafe)field.get(null);
}
catch (Exception ex) {
throw new RuntimeException("can't get Unsafe instance", ex);
}
}
public final static void main(String[] args) {
Unsafe u = getUnsafe();
Random r = new Random();
for (int i=0; i < 1000000; i++) {
u.setMemory(r.nextLong(), 1L, (byte)0x0);
}
}
}
#
# An unexpected error has been detected by Java Runtime Environment:
#
# SIGSEGV (0xb) at pc=0xb7df628a, pid=7635, tid=3084217232
#
# Java VM: Java HotSpot(TM) Client VM (11.0-b16 mixed mode, sharing linux-x86)
# Problematic frame:
# C [libc.so.6+0x7528a] memset+0x1a
#
