欢迎访问 生活随笔!

ag凯发k8国际

当前位置: ag凯发k8国际 > 编程语言 > c/c >内容正文

c/c

【c/c 13】天气app:数据挖掘/http协议/非结构化数据存储(filetoblob.cpp),数据管理/监控告警(hsmtable.cpp,tbspaceinfo.cpp) -ag凯发k8国际

发布时间:2024/4/24 c/c 186 豆豆
ag凯发k8国际 收集整理的这篇文章主要介绍了 【c/c 13】天气app:数据挖掘/http协议/非结构化数据存储(filetoblob.cpp),数据管理/监控告警(hsmtable.cpp,tbspaceinfo.cpp) 小编觉得挺不错的,现在分享给大家,帮大家做个参考.

文章目录

  • 1. 数据挖掘:/etc/rc.local,sudo su
  • 2.http协议:优先wget
  • 3.非结构化数据存储:blob,pzhrain24file
  • 4.数据管理子系统:数据字典表
  • 5.监控告警子系统:tmpfs,top,/proc/stat,/proc/meminfo,mount ,processes
    • 5.1 du,df,fdisk,lsblk:磁盘=硬盘
    • 5.2 文件:建表入表,传输

数据挖掘dmining和数据交换exptables差不多(exptables自己设计数据表有keyid字段,从其他地方拿数据未必有keyid字段需要指定,也就是数据是增量的,但不提供增量提取方式即没有keyid字段),将数据从数据库中取出导入文件(执行一个sql拿出数据)。vi /etc/rc.local(随操作系统自启动)。

vi /htidc/gz…sh上面如下,如下要写的其实就是< selectsql >字段里内容。比如取近一小时数据,每次取到都有重复,但在入库时(获取到的数据生成xml再入自己库)处理重复就行。

//dminoracle.cpp #include "_public.h" #include "_ooci.h" //oracle数据库是_ooci.h,mysql数据库换成_mysql.h,postgresql数据库换成_postgresql.h //公安内外网协议不通,数据挖掘不一定有数据,所以写入文件。 struct st_arg // 主程序的参数 {char connstr[101];char charset[51];char tname[51];char cols[1001];char fieldname[1001];char fieldlen[501];int exptype;char andstr[501];char bname[51];char ename[51];char idfieldname[51];char idfilename[301];char exppath[301];int timetvl; } starg; clogfile logfile; connection conn; bool _dmintables(); void exit(int sig); vector<string> vfieldname; // 存放拆分fieldname后的容器 vector<int> vfieldlen; // 存放拆分fieldlen后的容器 int maxfieldlen; // 存放fieldlen的最大值 void splitfields(); // 拆分fieldname和fieldlen void _help(char *argv[]); long maxkeyid; // 已挖掘数据的最大的id bool loadmaxkeyid(); // 从记录已获取数据最大id的文件中加载已挖掘数据的最大的id bool uptmaxkeyid(); // 更新已挖掘数据的最大的id到文件中 bool _xmltoarg(char *strxmlbuffer);int main(int argc,char *argv[]) {if (argc!=3) { _help(argv); return -1; }closeioandsignal();signal(sigint,exit); signal(sigterm,exit);if (logfile.open(argv[1],"a ")==false){printf("打开日志文件失败(%s)。\n",argv[1]); return -1;} if (_xmltoarg(argv[2])==false) return -1; //把xml解析到参数starg结构中while (true){if (conn.connecttodb(starg.connstr,starg.charset) != 0){logfile.write("connect database %s failed.\n",starg.connstr); sleep(starg.timetvl); continue;}// logfile.write("export table %s.\n",starg.tname);if (_dmintables() == false) logfile.write("export tables failed.\n");conn.disconnect(); // 断开与数据库的连接sleep(starg.timetvl);}return 0; } void exit(int sig) {logfile.write("程序退出,sig=%d\n\n",sig);exit(0); }//11111111111111111111111111111111111111111111显示程序的帮助 void _help(char *argv[]) {printf("\n");printf("using:/htidc/public/bin/dminoracle logfilename xmlbuffer\n\n");printf("增量挖掘示例:\n");printf("sample:/htidc/public/bin/dminoracle /log/shqx/dminoracle_surfdata_from_qx.log \"shqx/pwdidc@snorcl11g_198simplified chinese_china.zhs16gbkt_surfdataobtid,to_char(ddatetime,'yyyymmddhh24miss'),t,p,u,wd,wf,r,visobtid,ddatetime,t,p,u,wd,wf,r,vis5,14,8,8,8,8,8,8,81 and obtid in ('59293','50745')surfdata__from_qx/data/dmin/surfdata_from_qx.txtkeyid/data/shqx/sdata/fromqx30\"\n\n");printf("全量挖掘示例:\n");printf("sample:/htidc/public/bin/dminoracle /log/shqx/dminoracle_obtcode_from_qx.log \"shqx/pwdidc@snorcl11g_198simplified chinese_china.zhs16gbkt_obtcodeobtid,obtname,provname,lat,lon,heightobtid,obtname,provname,lat,lon,height5,30,30,8,8,82 and rsts=1 and obtid in ('59293','50745')obtcode__from_qx/data/shqx/sdata/fromqx300\"\n\n");printf("本程序是数据中心的公共功能模块,从其它业务系统的数据库中挖掘数据,用于入库到数据中心。\n");printf("logfilename是本程序运行的日志文件。\n");printf("xmlbuffer为文件传输的参数,如下:\n");printf("数据库的连接参数 shqx/pwdidc@snorcl11g_198\n");printf("数据库的字符集 simplified chinese_china.zhs16gbk 这个参数要与数据源数据库保持>一致,否则会出现中文乱码的情况。\n");printf("待挖掘数据的表名 t_surfdata\n");printf("需要挖掘字段的列表 obtid,to_char(ddatetime,'yyyymmddhh24miss'),t,p,u,wd,wf,r,vis 可以采用函数。\n");printf("挖掘字段的别名列表 obtid,ddatetime,t,p,u,wd,wf,r,vis 必须与cols一一对应。\n");printf("挖掘字段的长度列表 5,14,8,8,8,8,8,8,8 必须与cols一一对应。\n");printf("挖掘数据的方式 1 1-增量挖掘;2-全量挖掘,如果是增量挖掘,要求表一定要有表达记录序号的id字段。\n");printf("挖掘数据的附加条件 and obtid in ('59293','50745') 注意,关键字and不能少。\n");printf("数据文件的命名的前部分 surfdata_\n");printf("数据文件的命名的后部分 _from_qx\n");printf("挖掘数据表记录号字段名 keyid 当exptype=1时该参数有效。\n");printf("已挖掘数据id保存的文件名 /data/dmin/surfdata_from_qx.txt 当exptype=1时该参数有效。\n");printf("挖掘文件存放的目录 /data/shqx/sdata/fromqx\n");printf("挖掘数据的时间间隔 30 单位:秒,建议大于10。\n");printf("以上参数,除了idfieldname、idfilename和andstr,其它字段都不允许为空。\n\n\n"); }//1111111111111111111111111111111111111111111111把xml解析到参数starg结构中 bool _xmltoarg(char *strxmlbuffer) {memset(&starg,0,sizeof(struct st_arg));getxmlbuffer(strxmlbuffer,"connstr",starg.connstr);if (strlen(starg.connstr)==0) { logfile.write("connstr is null.\n"); return false; }getxmlbuffer(strxmlbuffer,"charset",starg.charset);if (strlen(starg.charset)==0) { logfile.write("charset is null.\n"); return false; }getxmlbuffer(strxmlbuffer,"tname",starg.tname);if (strlen(starg.tname)==0) { logfile.write("tname is null.\n"); return false; }getxmlbuffer(strxmlbuffer,"cols",starg.cols);if (strlen(starg.cols)==0) { logfile.write("cols is null.\n"); return false; }getxmlbuffer(strxmlbuffer,"fieldname",starg.fieldname);if (strlen(starg.fieldname)==0) { logfile.write("fieldname is null.\n"); return false; }getxmlbuffer(strxmlbuffer,"fieldlen",starg.fieldlen);if (strlen(starg.fieldlen)==0) { logfile.write("fieldlen is null.\n"); return false; }getxmlbuffer(strxmlbuffer,"exptype",&starg.exptype);if ( (starg.exptype!=1) && (starg.exptype!=2) ) { logfile.write("exptype is not in (1,2).\n"); return false; }getxmlbuffer(strxmlbuffer,"andstr",starg.andstr);if (strlen(starg.andstr)==0) { logfile.write("andstr is null.\n"); return false; }getxmlbuffer(strxmlbuffer,"bname",starg.bname);if (strlen(starg.bname)==0) { logfile.write("bname is null.\n"); return false; }getxmlbuffer(strxmlbuffer,"ename",starg.ename);if (strlen(starg.ename)==0) { logfile.write("ename is null.\n"); return false; }getxmlbuffer(strxmlbuffer,"idfieldname",starg.idfieldname);if ( (starg.exptype==1) && (strlen(starg.idfieldname)==0) ) { logfile.write("idfieldname is null.\n"); return false; }getxmlbuffer(strxmlbuffer,"idfilename",starg.idfilename);if ( (starg.exptype==1) && (strlen(starg.idfilename)==0) ) { logfile.write("idfilename is null.\n"); return false; }getxmlbuffer(strxmlbuffer,"exppath",starg.exppath);if (strlen(starg.exppath)==0) { logfile.write("exppath is null.\n"); return false; }getxmlbuffer(strxmlbuffer,"timetvl",&starg.timetvl);if (starg.timetvl==0) { logfile.write("timetvl is null.\n"); return false; }// 拆分fieldname和fieldlensplitfields();// 判断fieldname和fieldlen中元素的个数一定要相同if (vfieldname.size() != vfieldlen.size() ) { logfile.write("fieldname和fieldlen的元素个数不同。.\n"); return false; }return true; }//11111111111111111111111111111111111本程序的业务流程主函数 bool _dmintables() {// 从记录已获取数据最大id的文件中加载已挖掘数据的最大的idif (loadmaxkeyid()==false) { logfile.write("loadmaxkeyid() failed.\n"); return false; }// 生成挖掘数据的sql语句char strsql[4096]; char fieldvalue[vfieldname.size()][maxfieldlen1]; // 输出变量定义为一个二维数组memset(strsql,0,sizeof(strsql));if (starg.exptype==1)sprintf(strsql,"select %s,%s from %s where 1=1 and %s>%ld %s order by %s",starg.cols,starg.idfieldname,starg.tname,starg.idfieldname,maxkeyid,starg.andstr,starg.idfieldname);elsesprintf(strsql,"select %s from %s where 1=1 %s",starg.cols,starg.tname,starg.andstr);sqlstatement stmt(&conn);stmt.prepare(strsql);for (int ii=0;ii<vfieldname.size();ii){stmt.bindout(ii1,fieldvalue[ii],vfieldlen[ii]);} if (starg.exptype==1) stmt.bindout(vfieldname.size()1,&maxkeyid); //如果是增量挖掘,还要绑定id字段if (stmt.execute() != 0) // 执行挖掘数据的sql{logfile.write("select %s failed.\n%s\n%s\n",starg.tname,stmt.m_cda.message,stmt.m_sql); return false;}int ifileseq=1; // 待生成文件的序号char strfilename[301],strlocaltime[21];cfile file;while (true){memset(fieldvalue,0,sizeof(fieldvalue)); if (stmt.next() !=0) break; if (file.isopened()==false) // 把数据写入文件{memset(strlocaltime,0,sizeof(strlocaltime));localtime(strlocaltime,"yyyymmddhh24miss");memset(strfilename,0,sizeof(strfilename));sprintf(strfilename,"%s/%s%s%s_%d.xml",starg.exppath,starg.bname,strlocaltime,starg.ename,ifileseq);if (file.openforrename(strfilename,"w")==false){logfile.write("file.openforrename(%s) failed.\n",strfilename); return false;}file.fprintf("\n");}for (int ii=0;ii<vfieldname.size();ii){file.fprintf("<%s>%s",vfieldname[ii].c_str(),fieldvalue[ii],vfieldname[ii].c_str());}file.fprintf("\n");//1000条记录写入一个文件完成if (stmt.m_cda.rpc%1000==0) {file.fprintf("\n");if (file.closeandrename()==false){logfile.write("file.closeandrename(%s) failed.\n",strfilename); return false;}// 更新已挖掘数据的最大的id到文件中if (uptmaxkeyid()==false) { logfile.write("uptmaxkeyid() failed.\n"); return false; }logfile.write("create file %s ok.\n",strfilename);}}//不够1000条的写入一个文件if (file.isopened()==true){file.fprintf("\n");if (file.closeandrename()==false){logfile.write("file.closeandrename(%s) failed.\n",strfilename); return false;}// 更新已挖掘数据的最大的id到文件中if (uptmaxkeyid()==false) { logfile.write("uptmaxkeyid() failed.\n"); return false; }logfile.write("create file %s ok.\n",strfilename);}if (stmt.m_cda.rpc>0) logfile.write("本次挖掘了%d条记录。\n",stmt.m_cda.rpc);return true; }//111111111111111111111111111111111111拆分fieldname和fieldlen void splitfields() {vfieldname.clear(); vfieldlen.clear(); maxfieldlen=0; ccmdstr cmdstr;cmdstr.splittocmd(starg.fieldname,",");vfieldname.swap(cmdstr.m_vcmdstr);int ifieldlen=0;cmdstr.splittocmd(starg.fieldlen,",");for (int ii=0;ii<cmdstr.cmdcount();ii){ cmdstr.getvalue(ii,&ifieldlen);if (ifieldlen>maxfieldlen) maxfieldlen=ifieldlen; // 得到fieldlen的最大值vfieldlen.push_back(ifieldlen);} }//111111111111111111111111111111从记录已获取数据最大id的文件中加载已挖掘数据的最大的id bool loadmaxkeyid() {if (starg.exptype!=1) return true;cfile file;if (file.open(starg.idfilename,"r")==false) {logfile.write("注意,%s文件不存在,程序将从新开始挖掘数据。\n",starg.idfilename); return true;}char strbuf[21];memset(strbuf,0,sizeof(strbuf));file.fread(strbuf,20);maxkeyid=atol(strbuf);logfile.write("maxkeyid=%d\n",maxkeyid);return true; }//11111111111111111111111111111111更新已挖掘数据的最大的id到文件中 bool uptmaxkeyid() {if (starg.exptype!=1) return true;cfile file;if (file.open(starg.idfilename,"w")==false) {logfile.write("file.open(%s) failed.\n",starg.idfilename); return false;}file.fprintf("%ld",maxkeyid);return true; }

数据共享平台也有一个web系统(服务端)用java做的,vi /etc/rc.local如下。

如下http客户端访问不了,startup后可访问。

按照http格式向网站发报文也会得到回应,如下http客户端。


wgetclient是用wget命令下载,httpclient下载的内容有时候会有断行,wget下载不会。优先wget,wget下载不了的用http。wget支持http,https,ftp,sftp等协议。linux下同步文件rsync,curl,wget,ftp,sftp。

如上调用接口,如下还可以右击网页上图片复制图片地址。

程序在后台跑,有新图就拿下来,wgetclient将网页内容全弄下来,搞清图片命名规律进行解析。

//wgetclient.cpp #include "_public.h" void exit(int sig); clogfile logfile; int main(int argc, char *argv[]) {if(argc!=6){printf("usage:%s weburl tmpfilename outputfilename logfilename charset\n",argv[0]); printf("本程序用于获取web网页的内容。\n");printf("weburl 网页web的地址。\n");printf("tmpfilename 获取到的网页的内容存放的全路径的临时文件名,该文件可能是utf-8或其它编码。\n");printf("outputfilename 最终的输出文件全路径文件名,该文件是gb18030编码,注意tmpfilename被转换为outputfilename后,tmpfilename文件被自动删除。\n");printf("logfilename 本程序的运行产生的日志文件名。\n");printf("charset 网页的字符集,如utf-8\n\n");exit(1);}// 关闭全部的信号和输入输出// 设置信号,在shell状态下可用 "kill 进程号" 正常终止些进程// 但请不要用 "kill -9 进程号" 强行终止closeioandsignal(); signal(sigint,exit); signal(sigterm,exit);if (logfile.open(argv[4],"a ") == false){printf("logfile.open(%s) failed.\n",argv[4]); return -1;}mkdir(argv[2],true); mkdir(argv[3],true);char strweburl[3001];memset(strweburl,0,sizeof(strweburl));strncpy(strweburl,argv[1],3000);char strcmd[3024];memset(strcmd,0,sizeof(strcmd));snprintf(strcmd,3000,"/usr/bin/wget -c -q -o %s \"%s\" 1>>/dev/null 2>>/dev/null",argv[2],strweburl);system(strcmd);logfile.write("%s\n",strcmd);char strfilenametmp[301];memset(strfilenametmp,0,sizeof(strfilenametmp));snprintf(strfilenametmp,300,"%s.tmp",argv[3]);// 把获取到的网页转换为中文memset(strcmd,0,sizeof(strcmd));snprintf(strcmd,256,"iconv -c -f %s -t gb18030 %s -o %s",argv[5],argv[2],strfilenametmp);system(strcmd);logfile.write("%s\n",strcmd);remove(argv[2]); // 删除临时文件 rename(strfilenametmp,argv[3]);return 0; }void exit(int sig) {if (sig > 0) signal(sig,sig_ign);logfile.write("catching the signal(%d).\n",sig);logfile.write("wgetclient exit.\n");exit(0); } //wgetrain24.cpp #include "_public.h" void exit(int sig); clogfile logfile; bool geturl(char *strbuffer,char *strurl,char *strfilename); int main(int argc, char *argv[]) {if(argc!=4){printf("usage:%s logfilename tmpfilename outputfilename\n",argv[0]); printf("sample:./wgetrain24 /log/shqx/wgetrain24.log /data/wgettmp /data/wfile/zhrain24\n\n");printf("本程序用于从中国天气网获取逐小时降雨量实况图。\n");printf("中国天气网的url是http://products.weather.com.cn/product/index/index/procode/jc_jsl_zh.shtml\n");printf("如果中国天气网的url改变,程序也在做改动。\n");printf("logfilename 本程序的运行产生的日志文件名。\n");printf("tmpfilename 本程序运行产生的临时文件存放的目录。\n");printf("获取逐小时降雨量实况图存放的目录。\n\n");exit(1);}closeioandsignal(); signal(sigint,exit); signal(sigterm,exit);if (logfile.open(argv[1],"a ") == false){printf("logfile.open(%s) failed.\n",argv[1]); return -1;}mkdir(argv[2],false); mkdir(argv[3],false);while (true){ char strwgetclient[2001]; // 调用wgetclient获取网页内容memset(strwgetclient,0,sizeof(strwgetclient));snprintf(strwgetclient,2000,"/htidc/public/bin/wgetclient \"http://products.weather.com.cn/product/index/index/procode/jc_jsl_zh.shtml\" %s/wgetclient_%d.tmp %s/wgetclient_%d.html %s/wgetclient.log utf-8",argv[2],getpid(),argv[2],getpid(),argv[2]);system(strwgetclient);// logfile.write("%s\n",strwgetclient);// 打开网页内容文件char stroutputfile[301];memset(stroutputfile,0,sizeof(stroutputfile));snprintf(stroutputfile,300,"%s/wgetclient_%d.html",argv[2],getpid());cfile file;if (file.open(stroutputfile,"r")==false){logfile.write("file.open(%s) failed.\n",stroutputfile); sleep(60); continue;} char strbuffer[1001],strurl[501],strfullfilename[301],strfilename[101]; // 得到全部的图片文件名while (true){memset(strbuffer,0,sizeof(strbuffer));memset(strurl,0,sizeof(strurl));memset(strfullfilename,0,sizeof(strfullfilename));memset(strfilename,0,sizeof(strfilename)); if (file.fgets(strbuffer,1000)==false) break; if (matchfilename(strbuffer,"*pwcp_twc_weap_sfer_er1_twc_l88_p9_20*.jpg*")==false) continue; // logfile.write("%s",strbuffer); if (geturl(strbuffer,strurl,strfilename)==false) continue; //解析出url和文件名 snprintf(strfullfilename,300,"%s/%s",argv[3],strfilename); //文件已存在,不采集if (access(strfullfilename,f_ok)==0) continue; logfile.write("download %s ",strfilename); //调用wget获取文件memset(strwgetclient,0,sizeof(strwgetclient));snprintf(strwgetclient,500,"wget \"%s\" -o %s/wgetrain24.log -o %s",strurl,argv[2],strfullfilename);system(strwgetclient); if (access(strfullfilename,f_ok)==0) logfile.writeex("ok.\n");else logfile.writeex("failed.\n");} file.closeandremove();sleep(60);}return 0; }bool geturl(char *strbuffer,char *strurl,char *strfilename) {char *start,*end;start=end=0;if ((start=strstr(strbuffer,"http"))==0) return false;if ((end=strstr(start,"\""))==0) return false; //找双引号strncpy(strurl,start,end-start);strcpy(strfilename,strstr(strurl,"pwcp"));return true; } void exit(int sig) {if (sig > 0) signal(sig,sig_ign);logfile.write("catching the signal(%d).\n",sig);logfile.write("wgetclient exit.\n");exit(0); } // 本程序演示如何把磁盘文件的文本文件写入oracle的blob字段中。 //filetoblob.cpp,实时生成的不要存oracle的blob字段 #include "_ooci.h" int main(int argc,char *argv[]) {connection conn; // 连接数据库,返回值0-成功,其它-失败// 失败代码在conn.m_cda.rc中,失败描述在conn.m_cda.message中。if (conn.connecttodb("scott/tiger@snorcl11g_198","simplified chinese_china.zhs16gbk") != 0){printf("connect database %s failed.\n%s\n","scott/tiger@orcl",conn.m_cda.message); return -1;} sqlstatement stmt(&conn); // sql语言操作类 // 为了方便演示,把goods表中的记录全删掉,再插入一条用于测试的数据。// 不需要判断返回值stmt.prepare("\begin\delete from goods;\insert into goods(id,name,pic) values(1,'商品名称',empty_blob());\end;"); // 执行sql语句,一定要判断返回值,0-成功,其它-失败。if (stmt.execute() != 0){printf("stmt.execute() failed.\n%s\n%s\n",stmt.m_sql,stmt.m_cda.message); return -1;}// 使用游标从goods表中提取id为1的pic字段// 注意了,同一个sqlstatement可以多次使用// 但是,如果它的sql改变了,就要重新prepare和bindin或bindout变量stmt.prepare("select pic from goods where id=1 for update");stmt.bindblob();// 执行sql语句,一定要判断返回值,0-成功,其它-失败。if (stmt.execute() != 0){printf("stmt.execute() failed.\n%s\n%s\n",stmt.m_sql,stmt.m_cda.message); return -1;}// 获取一条记录,一定要判断返回值,0-成功,1403-无记录,其它-失败。if (stmt.next() != 0) return 0; // 把磁盘文件pic_in.jpg的内容写入blob字段,一定要判断返回值,0-成功,其它-失败。if (stmt.filetolob((char *)"pic_in.jpg") != 0){printf("stmt.filetolob() failed.\n%s\n",stmt.m_cda.message); return -1;} conn.commit(); // 提交事务return 0; }




// pzhrain24file.cpp #include "_public.h" #include "_ooci.h" #include "_shqx.h" clogfile logfile; cdir dir; // 处理数据文件 bool _pzhrain24file(char *strargv2,char *strargv4,char *strargv5); connection conn; void exit(int sig);int main(int argc,char *argv[]) {if (argc!=7){printf("\n本程序用于处理全国逐小时雨量实况图片文件。\n\n");printf("/htidc/shqx/bin/pzhrain24file logfilename connstr srcpathname dstpathname tname timetvl\n");printf("例如:/htidc/shqx/bin/pzhrain24file /log/shqx/pzhrain24file.log shqx/pwdidc@snorcl11g_198 /data/wfile/zhrain24 /qxfile/zhrain24 t_zhrain24 30\n");printf("logfilename 本程序运行的日志文件名。\n");printf("connstr 数据库的连接参数。\n");printf("srcpathname 原始文件存放的目录,文件命名如pwcp_twc_weap_sfer_er1_twc_l88_p9_20191101070000000.jpg。\n");printf("dstpathname 目标文件存放的目录,文件按yyyy/mm/dd组织目录,重命名为zhrain24_yyyymmddhh24miss.jpg。\n");printf("tname 数据存放的表名。\n");printf("timetvl 本程序运行的时间间隔,单位:秒。\n");return -1;}closeioandsignal();signal(sigint,exit); signal(sigterm,exit);if (logfile.open(argv[1],"a ")==false){printf("打开日志文件失败(%s)。\n",argv[1]); return -1;}logfile.write("程序启动。\n");while (true){// logfile.write("开始扫描目录。\n");// 扫描数据文件存放的目录,只匹配"pwcp_twc_weap_sfer_er1_twc_l88_p9_20*.jpg"if (dir.opendir(argv[3],"pwcp_twc_weap_sfer_er1_twc_l88_p9_20*.jpg",1000,true,true)==false){logfile.write("dir.opendir(%s) failed.\n",argv[3]); sleep(atoi(argv[6])); continue;} while (true) // 逐个处理目录中的数据文件{if (dir.readdir()==false) break;if (_pzhrain24file(argv[2],argv[4],argv[5])==false) {logfile.writeex("失败。\n"); continue;}}if (conn.m_state==1) conn.disconnect(); sleep(atoi(argv[6]));}return 0; } void exit(int sig) {logfile.write("程序退出,sig=%d\n\n",sig);exit(0); }//1111111111111111111111111111111111111111111处理数据文件 bool _pzhrain24file(char *strargv2,char *strargv4,char *strargv5) {char strddatetime[21]; // 文件的数据时间,格式yyyymmddhh24missmemset(strddatetime,0,sizeof(strddatetime));strncpy(strddatetime,strstr(dir.m_filename,"20"),14);//搜索文件名pwcp_twc…中20,后面取14位,重命名为zhrain24_%s.jpgchar strdstfilename[301]; // 目标文件名,不带路径memset(strdstfilename,0,sizeof(strdstfilename));snprintf(strdstfilename,300,"zhrain24_%s.jpg",strddatetime);char strdstfilepath[301]; // 目标文件存放的目录memset(strdstfilepath,0,sizeof(strdstfilepath));snprintf(strdstfilepath,300,"%s/",strargv4);strncat(strdstfilepath,strddatetime,4); strcat(strdstfilepath,"/"); // 年的子目录strncat(strdstfilepath,strddatetime4,2); strcat(strdstfilepath,"/"); // 月的子目录strncat(strdstfilepath,strddatetime6,2); strcat(strdstfilepath,"/"); // 日的子目录char strfulldstfilename[301]; // 目标文件名,全路径memset(strfulldstfilename,0,sizeof(strfulldstfilename));snprintf(strfulldstfilename,300,"%s%s",strdstfilepath,strdstfilename);// 如果文件已处理(目标文件已存在),直接返回成功。if (access(strfulldstfilename,f_ok) == 0) return true;if (conn.m_state==0){if (conn.connecttodb(strargv2,"simplified chinese_china.zhs16gbk")!=0){logfile.write("connect database(%s) failed.\n%s\n",strargv2,conn.m_cda.message); return false;}// logfile.write("连接数据库成功。\n");}// 把源文件复制到目标文件if (copy(dir.m_fullfilename,strfulldstfilename)==false) {logfile.write("复制文件copy(%s,%s)...failed.\n",dir.m_fullfilename,strfulldstfilename); return false;}// 把非结构化数据文件写入oracle数据库的表中if (filetotable(&conn,&logfile,strargv5,strfulldstfilename,strddatetime)!=0){logfile.write("把文件%s存入%s...failed.\n",strfulldstfilename,strargv5); return false;}logfile.write("把文件%s存入%s...ok.\n",strfulldstfilename,strargv5); return true; }



oracle数据字典:是一组表和视图的结构(就像仓库里有什么,每种有多少,哪些空间可用等日记)。自己创建的表,oracle会将这表的信息自动写入系统表,系统表上再创建一视图供客户查看。数据字典中的表是不能被访问修改(自己用的),但是可以访问数据字典的视图就可以知道自己创建的表的详细信息。


tab也是视图,tab里面表,视图,同义词都可以查到。若是select * from dba_tables;必须是dba用户才可以查,三类user_,all_,dba_

数据量大了会出现性能和数据管理,迁移,备份问题。假如有1亿数据量,不能用exp导出,一个文件有几百g导出要一天。索引设计要合理,不然table scan跑不动。一亿以下的数据单表存放,一亿以上的数据考虑分表,十亿虽然比一亿容量大十倍,但性能不会下降十倍,因为索引(比如按首字母找名字)。

如下并行数据库分布式查询可解决多表性能低的问题。

写一些程序实现数据管理,数据删除(deletables.cpp)或放历史表等等。三种表数据结构一样,迁移数据可写成一个通用功能模块程序。若是不同表,只知道表名去查询数据字典得到数据结构。如下hsmtable.cpp(和deletetable.cpp像):数据源表就是要把数据从哪个表迁移出来,列名从数据字典取出,把列拼成一个字符串,获取列名后把数据rowid从源表查出来,生成插入目的表sql,删除源表sql。

//hsmtable.cpp #include "_public.h" #include "_ooci.h" char logfilename[301]; char connstr[101]; char srctname[51]; char dsttname[51]; char where[1024]; char hourstr[101]; char localhour[21]; int maxcount=1; connection conn; clogfile logfile; void exit(int sig); char strcolumnstr[2048]; //获取dsttname表全部的列 bool getcolumnstr(); void _help(char *argv[]); bool _hsmtables();int main(int argc,char *argv[]) {if (argc != 2) { _help(argv); return -1; }memset(logfilename,0,sizeof(logfilename));memset(connstr,0,sizeof(connstr));memset(srctname,0,sizeof(srctname));memset(dsttname,0,sizeof(dsttname));memset(where,0,sizeof(where));memset(hourstr,0,sizeof(hourstr));getxmlbuffer(argv[1],"logfilename",logfilename,300);getxmlbuffer(argv[1],"connstr",connstr,100);getxmlbuffer(argv[1],"srctname",srctname,50);getxmlbuffer(argv[1],"dsttname",dsttname,50);getxmlbuffer(argv[1],"where",where,1000);getxmlbuffer(argv[1],"maxcount",&maxcount);getxmlbuffer(argv[1],"hourstr",hourstr,2000);if (strlen(logfilename) == 0) { printf("logfilename is null.\n"); return -1; }if (strlen(connstr) == 0) { printf("connstr is null.\n"); return -1; }if (strlen(srctname) == 0) { printf("srctname is null.\n"); return -1; }if (strlen(dsttname) == 0) { printf("dsttname is null.\n"); return -1; }if (strlen(where) == 0) { printf("where is null.\n"); return -1; }if ( (maxcount<1) || (maxcount>1000) ) { printf("maxcount %d is invalid,should in 1-1000.\n",maxcount); return -1; }if (strlen(hourstr) == 0) { printf("hourstr is null.\n"); return -1; }closeioandsignal();signal(sigint,exit); signal(sigterm,exit);if (logfile.open(logfilename,"a ") == false){printf("logfile.open(%s) failed.\n",logfilename); return -1;}while (true){ memset(localhour,0,sizeof(localhour)); //判断当前时间是否在启动时间之内localtime(localhour,"hh24");if (strstr(hourstr,localhour)==0) { sleep(60); continue; }if (conn.connecttodb(connstr,"simplified chinese_china.zhs16gbk") != 0){logfile.write("connect database %s failed.\n",connstr); sleep(60); continue; }logfile.write("from table %s to %s.\n",srctname,dsttname);if (_hsmtables() == false) logfile.write("_hsmtables failed.\n"); conn.disconnect();sleep(60); }return 0; } void exit(int sig) {printf("程序退出,sig=%d\n\n",sig);exit(0); }//111111111111111111111111111111111111111111111111111111111111111 void _help(char *argv[]) {printf("\nusing:/htidc/public/bin/hsmtables \"/log/shqx/hsmtables_surfdata.logshqx/pwdidc@snorcl11g_198t_surfdatat_surfdata_hiswhere ddatetime50023,01,02,03,04,05,06\"\n\n");printf("这是一个工具程序,用于清理表中的数据。\n");printf("/log/shqx/hsmtables_surfdata.log 本程序运行日志文件名。\n");printf("szidc/pwdidc@szqx_10.153.97.251 目的数据库的连接参数。\n");printf("t_surfdata 数据源表名。\n");printf("t_surfdata_his 目的数据表名。\n");printf("where ddatetime 待迁移数据的条件。\n");printf("500 单次执行数据迁移的记录数,取值在1-1000之间。\n");printf("23,01,02,03,04,05,06 本程序启动的时次,小时,时次之间用半角的逗号分隔开。\n\n");return; }//11111111111111111111111111111111111111111111111111111111111111 bool _hsmtables() { if (getcolumnstr() == false) return false; //获取dsttname表全部的列int ccount=0;char strrowid[51],strrowidn[maxcount][51];sqlstatement selstmt(&conn); //把数据的rowid从srctname表中查出来selstmt.prepare("select rowid from %s %s",srctname,where);selstmt.bindout(1, strrowid,50);if (selstmt.execute() != 0){logfile.write("%s failed.\n%s\n",selstmt.m_sql,selstmt.m_cda.message); return false;}int ii=0; // 生成插入dsttname表和删除srctname表的sqlchar strinsertsql[10241],strdeletesql[10241];memset(strinsertsql,0,sizeof(strinsertsql));memset(strdeletesql,0,sizeof(strdeletesql));sprintf(strinsertsql,"insert into %s(%s) select %s from %s where rowid in (",dsttname,strcolumnstr,strcolumnstr,srctname);//%s(%s)里面的%s是字段名,字段顺序没必要保持一致,所有把列名从数据字典取出,目的表字段比原表少,只备份一部分字段sprintf(strdeletesql,"delete from %s where rowid in (",srctname);char strtemp[11];for (ii=0; ii<maxcount; ii){memset(strtemp,0,sizeof(strtemp));if (ii==0) sprintf(strtemp,":%d",ii1);if (ii >0) sprintf(strtemp,",:%d",ii1);strcat(strinsertsql,strtemp);strcat(strdeletesql,strtemp);}strcat(strinsertsql,")");strcat(strdeletesql,")");sqlstatement insstmt(&conn);insstmt.prepare(strinsertsql);sqlstatement delstmt(&conn);delstmt.prepare(strdeletesql);for (ii=0; ii<maxcount; ii){insstmt.bindin(ii1,strrowidn[ii],50);delstmt.bindin(ii1,strrowidn[ii],50);}while (true) //每maxcount记录就执行一次{memset(strrowid,0,sizeof(strrowid));if (selstmt.next() != 0) break;strcpy(strrowidn[ccount],strrowid);ccount;if (ccount == maxcount){if (insstmt.execute() != 0){if (insstmt.m_cda.rc != 1){logfile.write("_hsmtables insert %s failed.\n%s\n",dsttname,insstmt.m_cda.message); return false;}}if (delstmt.execute() != 0){logfile.write("_hsmtables delete %s failed.\n%s\n",dsttname,insstmt.m_cda.message); return false;}conn.commit();memset(strrowidn,0,sizeof(strrowidn));ccount=0;}if (fmod(selstmt.m_cda.rpc,10000) < 1){logfile.write("%s to %s ok(%d).\n",srctname,dsttname,selstmt.m_cda.rpc); memset(localhour,0,sizeof(localhour)); // 判断当前时间是否在启动时间之内localtime(localhour,"hh24");if (strstr(hourstr,localhour)==0) return true;}}// 在以上循环处理的时候,如果不足maxcount,就在这里处理for (ii=0; ii<ccount; ii){insstmt.prepare("\begin\insert into %s(%s) select %s from %s where rowid=:1;\delete from %s where rowid=:2;\end;",dsttname,strcolumnstr,strcolumnstr,srctname,srctname);insstmt.bindin(1,strrowidn[ii],50);insstmt.bindin(2,strrowidn[ii],50);if (insstmt.execute() != 0){if (insstmt.m_cda.rc != 1){logfile.write("_hsmtables insert %s or delete %s failed.\n%s\n",dsttname,srctname,insstmt.m_cda.message); return false;}}}conn.commit(); //一行一行的提交logfile.write("%s to %s finish(%d).\n",srctname,dsttname,selstmt.m_cda.rpc);return true; }//1111111111111111111111111111111111获取dsttname表全部的列 bool getcolumnstr() {memset(strcolumnstr,0,sizeof(strcolumnstr)); char column_name[51];sqlstatement stmt(&conn);stmt.prepare("select lower(column_name) from user_tab_columns where table_name=upper('%s') order by column_id",dsttname);stmt.bindout(1,column_name,50);if (stmt.execute() != 0){logfile.write("%s failed.\n%s\n",stmt.m_sql,stmt.m_cda.message); return false;}while(true){memset(column_name,0,sizeof(column_name));if (stmt.next()!=0) break;if (stmt.m_cda.rpc>1) strcat(strcolumnstr,",");strcat(strcolumnstr,column_name);}if (stmt.m_cda.rpc==0) { logfile.write("表%s不存在。\n",dsttname); return false; }return true; }


$sqlplus shqx/pwdidc,如下数据都是一小时前生成,需要将后台生成数据脚本启动。

如下是数据迁移程序如何兼容其他数据库,oracle有rowid,mysql和pg都没有rowid但有keyid。

不管id字段是数字还是字符或字符串,我们都可用字符绑定。

如下不同数据库查数据字典得到全部列。


我们处理后台数据也会有一个web系统用来管理参数,展示数据,查询数据等界面(不能让用户去登录plsqldeveloper查询)。如下是数据的监控,数据少于某数量显示红色。


如下日志等等都在sda(raid1,2个300g)中。

5.1 du,df,fdisk,lsblk:磁盘=硬盘

du:disk usage。-a:不仅统计目录还包括文件大小。-h:显示k/m/g等字节信息。-s:总的占用空间大小(-a是每个都统计)。一般文件夹大小4096b即4k(不包含里面文件大小)。


df:disk free。如果你想知道某个文件夹或文件大小用du,磁盘相关使用情况用df。





5.2 文件:建表入表,传输

以下是收集磁盘空间信息,现在插入一个u盘,fdisk -l可以看到u盘。



每个服务器上运行一个收集磁盘空间的小程序,收集到磁盘空间后生成xml文件存放在本地目录,通过文件传输系统或ftp将文件传给数据处理服务器(因为有的服务器不安装oracle客户端,只可安装文件传输客户端),统一保存到数据库。

//diskinfo.cpp,写入xml文件中再入库 #include "_public.h" void exit(int sig); clogfile logfile;int main(int argc,char *argv[]) {if (argc != 4){printf("\n");printf("using:./diskinfo hostname logfilename outputpath\n");printf("example:/htidc/public/bin/diskinfo 118.89.50.198 /tmp/htidc/log/diskinfo.log /tmp/htidc/monclient\n\n");printf("此程序调用df命名,把本服务器的磁盘使用率信息写入xml文件。\n");printf("hostname是本服务器的主机名,为了方便识别,也可以用ip。\n");printf("logfilename是本程序的日志文件名。\n");printf("outputpath是输出的xml文件存放的目录。\n");printf("此程序运行在需要监控的服务器上(本程序只适用linux系统),采集后的xml文件由文件传输程序发送给数据处理服务程序入库。\n\n\n");return -1;}closeioandsignal(); signal(sigint,exit); signal(sigterm,exit);if (logfile.open(argv[2],"a ") == false){printf("logfile.open(%s) failed.\n",argv[2]); return -1;}file *fp=0;if ( (fp=popen("df -k --block-size=1m","r")) == null ){logfile.write("popen(df -k --block-size=1m) failed.\n"); return false;}char strxmlfilename[301],strlocaltime[21];memset(strxmlfilename,0,sizeof(strxmlfilename));memset(strlocaltime,0,sizeof(strlocaltime));localtime(strlocaltime,"yyyymmddhh24miss");snprintf(strxmlfilename,300,"%s/diskinfo_%s_%s.xml",argv[3],strlocaltime,argv[1]);cfile xmlfile;if (xmlfile.openforrename(strxmlfilename,"w ") == false ){logfile.write("xmlfile.openforrename(%s) failed.\n",strxmlfilename); pclose(fp); return -1;}xmlfile.fprintf("\n");ccmdstr cmdstr;char strbuffer[1024],strline[500];while (true){memset(strbuffer,0,sizeof(strbuffer));if (fgets(fp,strbuffer,500) == false) break;// 如果没有找到“%”,就再读取一行,与strbuffer拼起来if (strstr(strbuffer,"%") == 0){memset(strline,0,sizeof(strline));if (fgets(fp,strline,500) == false) break;strcat(strbuffer," "); strcat(strbuffer,strline);}// 删除字符串前后的空格和换行符deletelrchar(strbuffer,' '); deletelrchar(strbuffer,'\n');// 把字符串中间的多个空格全部转换为一个空格updatestr(strbuffer," "," ");// 把全内容全部转换为小写tolower(strbuffer);// 除了磁盘信息,还有可能是内存,smb等其它文件,都丢弃掉if (strncmp(strbuffer,"/dev",4) != 0) continue;cmdstr.splittocmd(strbuffer," ");if (cmdstr.cmdcount() != 6) continue;char strusep[21];memset(strusep,0,sizeof(strusep));strcpy(strusep,cmdstr.m_vcmdstr[4].c_str());updatestr(strusep,"%","");char strlocaltime[21];memset(strlocaltime,0,sizeof(strlocaltime));localtime(strlocaltime,"yyyymmddhh24miss");xmlfile.fprintf(\"%s"\"%s"\"%s"\"%0.02f"\"%0.02f"\"%0.02f"\"%0.02f"\"%s\n",argv[1],strlocaltime,cmdstr.m_vcmdstr[0].c_str(),atof(cmdstr.m_vcmdstr[1].c_str())/1024.0,atof(cmdstr.m_vcmdstr[2].c_str())/1024.0,atof(cmdstr.m_vcmdstr[3].c_str())/1024.0,(atof(cmdstr.m_vcmdstr[2].c_str())/atof(cmdstr.m_vcmdstr[1].c_str()))*100.0,cmdstr.m_vcmdstr[5].c_str());}xmlfile.fprintf("\n");pclose(fp);xmlfile.closeandrename();logfile.write("create %s ok.\n",strxmlfilename);exit(0); }void exit(int sig) {if (sig > 0) signal(sig,sig_ign);logfile.write("catching the signal(%d).\n",sig);logfile.write("diskinfo exit.\n");exit(0); }

以下为收集cpu和内存信息,top命令显示如下,zombie表示僵尸,q退出,cpu用到20%算忙了。

ps -ef |grep htidc,cpuinfo.cpp思路是定义三个结构体变量,加载cpu信息到结构体里,睡60s,再继续加载cpu信息到结构体里,再将两结构体成员相减,就可以知道一分钟内cpu情况,采用的是一分钟信息。vi /proc/stat如下。

//cpuinfo.cpp #include "_public.h" void exit(int sig); clogfile logfile; struct st_cpuinfo {double user;double sys;double wait;double nice;double idle;double irq;double softirq;double total; }; struct st_cpuinfo stcpuinfo1,stcpuinfo2,stcpuinfo3; bool loadcpuinfo(struct st_cpuinfo &stcpuinfo);int main(int argc,char *argv[]) {if (argc != 4){printf("\n");printf("using:./cpuinfo hostname logfilename outputpath\n");printf("example:/htidc/public/bin/cpuinfo 118.89.50.198 /tmp/htidc/log/cpuinfo.log /tmp/htidc/monclient\n\n");printf("此程序读取/proc/stat文件,把本服务器的cpu使用率信息写入xml文件。\n");printf("hostname是本服务器的主机名,为了方便识别,也可以用ip。\n");printf("logfilename是本程序的日志文件名。\n");printf("outputpath是输出的xml文件存放的目录。\n");printf("此程序运行在需要监控的服务器上(本程序只适用linux系统),采集后的xml文件由文件传输程序发送给数据处理服务程序入库。\n\n\n");return -1;}//memset(strhostname,0,sizeof(strhostname));//strncpy(strhostname,argv[2],20);closeioandsignal(); signal(sigint,exit); signal(sigterm,exit);if (logfile.open(argv[2],"a ") == false){printf("logfile.open(%s) failed.\n",argv[2]); return -1;}memset(&stcpuinfo1,0,sizeof(struct st_cpuinfo));memset(&stcpuinfo2,0,sizeof(struct st_cpuinfo));memset(&stcpuinfo3,0,sizeof(struct st_cpuinfo));if (loadcpuinfo(stcpuinfo1) ==false) return -1; sleep(60);if (loadcpuinfo(stcpuinfo2) ==false) return -1;stcpuinfo3.user=stcpuinfo2.user-stcpuinfo1.user;stcpuinfo3.sys=stcpuinfo2.sys-stcpuinfo1.sys;stcpuinfo3.wait=stcpuinfo2.wait-stcpuinfo1.wait;stcpuinfo3.nice=stcpuinfo2.nice-stcpuinfo1.nice;stcpuinfo3.idle=stcpuinfo2.idle-stcpuinfo1.idle;stcpuinfo3.irq=stcpuinfo2.irq-stcpuinfo1.irq;stcpuinfo3.softirq=stcpuinfo2.softirq-stcpuinfo1.softirq;stcpuinfo3.total=stcpuinfo3.userstcpuinfo3.sysstcpuinfo3.waitstcpuinfo3.nicestcpuinfo3.idlestcpuinfo3.irqstcpuinfo3.softirq;char strlocaltime[21];memset(strlocaltime,0,sizeof(strlocaltime));localtime(strlocaltime,"yyyymmddhh24miss");char strxmlfilename[301];memset(strxmlfilename,0,sizeof(strxmlfilename));snprintf(strxmlfilename,300,"%s/cpuinfo_%s_%s.xml",argv[3],strlocaltime,argv[1]);cfile xmlfile;if (xmlfile.openforrename(strxmlfilename,"w ") == false ){logfile.write("xmlfile.openforrename(%s) failed.\n",strxmlfilename); return -1;}xmlfile.fprintf("\n");xmlfile.fprintf("%s%s%0.02f%0.02f%0.02f%0.02f%0.02f%0.02f\n",argv[1],strlocaltime,stcpuinfo3.user/stcpuinfo3.total*100.0,stcpuinfo3.sys/stcpuinfo3.total*100.0,stcpuinfo3.wait/stcpuinfo3.total*100.0,stcpuinfo3.nice/stcpuinfo3.total*100.0,stcpuinfo3.idle/stcpuinfo3.total*100.0,100.0-stcpuinfo3.nice/stcpuinfo3.total*100.0);xmlfile.fprintf("\n");xmlfile.closeandrename();logfile.write("create %s ok.\n",strxmlfilename);exit(0); } void exit(int sig) {if (sig > 0) signal(sig,sig_ign);logfile.write("catching the signal(%d).\n",sig);logfile.write("cpuinfo exit.\n");exit(0); }bool loadcpuinfo(struct st_cpuinfo &stcpuinfo) {cfile cpufile;if (cpufile.open("/proc/stat","r") == false ){logfile.write("cpufile.openforread(/proc/stat) failed.\n"); return false;}ccmdstr cmdstr;char strbuffer[1024];while (true){memset(strbuffer,0,sizeof(strbuffer));if (cpufile.ffgets(strbuffer,500) == false) break;// 删除字符串前后的空格deletelrchar(strbuffer,' ');// 把字符串中间的多个空格全部转换为一个空格updatestr(strbuffer," "," ");tolower(strbuffer);cmdstr.splittocmd(strbuffer," ");if (strcmp(cmdstr.m_vcmdstr[0].c_str(),"cpu")==0) {stcpuinfo.user=atof(cmdstr.m_vcmdstr[1].c_str());stcpuinfo.sys=atof(cmdstr.m_vcmdstr[2].c_str());stcpuinfo.wait=atof(cmdstr.m_vcmdstr[3].c_str());stcpuinfo.nice=atof(cmdstr.m_vcmdstr[4].c_str());stcpuinfo.idle=atof(cmdstr.m_vcmdstr[5].c_str());stcpuinfo.irq=atof(cmdstr.m_vcmdstr[6].c_str());stcpuinfo.softirq=atof(cmdstr.m_vcmdstr[7].c_str());return true;}}logfile.write("read /proc/stat failed.\n"); return false; }


vi /tmp/htdic/monclient/cpu*,如下是收集到的信息。

收集内存信息#free -m,和top命令查看的内存是一样的,也在系统文件vi /proc/meminfo。

以下是收集oracle表空间信息,表空间就像磁盘空间一样,表空间的信息收集要去读取oracle的数据字典,如下sql是查询oracle表空间使用率,取出后写入xml文件里,vi tbspaceinfo.cpp,要dba权限。

select * from ( select a.tablespace_name, to_char(a.bytes/1024/1024,'99,999.999') total_bytes, to_char(b.bytes/1024/1024,'99,999.999') free_bytes, to_char(a.bytes/1024/1024 - b.bytes/1024/1024,'99,999.999') use_bytes, to_char((1 - b.bytes/a.bytes)*100,'99.99') || '%' use from (select tablespace_name, sum(bytes) bytes from dba_data_files group by tablespace_name) a, (select tablespace_name, sum(bytes) bytes from dba_free_space group by tablespace_name) b where a.tablespace_name = b.tablespace_name union all select c.tablespace_name, to_char(c.bytes/1024/1024,'99,999.999') total_bytes, to_char( (c.bytes-d.bytes_used)/1024/1024,'99,999.999') free_bytes, to_char(d.bytes_used/1024/1024,'99,999.999') use_bytes, to_char(d.bytes_used*100/c.bytes,'99.99') || '%' use from (select tablespace_name,sum(bytes) bytes from dba_temp_files group by tablespace_name) c, (select tablespace_name,sum(bytes_cached) bytes_used from v$temp_extent_pool group by tablespace_name) d where c.tablespace_name = d.tablespace_name ) //tbspaceinfo.cpp #include "_public.h" #include "_ooci.h" void exit(int sig); struct st_tbspaceinfo {long taskid;char nodip[31];char tablespace[101];double total;double used;double available;double usep;double alarmvalue;int alarmsts;char crttime[21];int rsts; }; struct st_tbspaceinfo sttbspaceinfo; clogfile logfile; connection conn; int main(int argc,char *argv[]) {if (argc != 5){printf("\n");printf("using:./tbspaceinfo hostname logfilename outputpath username/password@tnsnames\n");printf("example:/htidc/public/bin/tbspaceinfo 10.153.98.13 /tmp/htidc/log/tbspaceinfo_10.153.98.13.log /tmp/htidc/monclient shqx/pwdidc@szqx_10.153.98.13\n\n");printf("此程序连接远程数据库,把远程数据库表空间使用率信息写入xml文件。\n");printf("hostname是本服务器的主机名,为了方便识别,也可以用ip。\n");printf("logfilename是本程序的日志文件名。\n");printf("outputpath是输出的xml文件存放的目录。\n");printf("username/password@tnsnames为待监控的远程数据库的用户名/密码@连接名。\n");printf("此程序运行在数据中心应用程序的服务器上。\n\n\n");return -1;}// 关闭全部的信号和输入输出// 设置信号,在shell状态下可用 "kill 进程号" 正常终止些进程// 但请不要用 "kill -9 进程号" 强行终止closeioandsignal(); signal(sigint,exit); signal(sigterm,exit);if (logfile.open(argv[2],"a ") == false){printf("logfile.open(%s) failed.\n",argv[2]); return -1;}if (conn.connecttodb(argv[4],"simplified chinese_china.zhs16gbk") != 0){logfile.write("conn.connecttodb(%s) failed.\n",argv[4]); return -1;}sqlstatement stmt;stmt.connect(&conn);stmt.prepare("\select f.tablespace_name,a.total,u.used,f.free,(u.used/a.total)*100 from\(select tablespace_name,sum(bytes/(1024*1024*1024)) total from dba_data_files\group by tablespace_name) a,\(select tablespace_name,round(sum(bytes/(1024*1024*1024))) used from dba_extents\group by tablespace_name) u,\(select tablespace_name,round(sum(bytes/(1024*1024*1024))) free from dba_free_space\group by tablespace_name) f\where a.tablespace_name = f.tablespace_name\and a.tablespace_name = u.tablespace_name\and f.tablespace_name in (select tablespace_name from dba_tablespaces where contents='permanent')");stmt.bindout(1,sttbspaceinfo.tablespace,100);stmt.bindout(2,&sttbspaceinfo.total);stmt.bindout(3,&sttbspaceinfo.used);stmt.bindout(4,&sttbspaceinfo.available);stmt.bindout(5,&sttbspaceinfo.usep);if (stmt.execute() != 0){logfile.write("select dba_data_files,dba_extents,dba_free_space failed.\n%s\n",stmt.m_cda.message); return -1;}char strlocaltime[21];memset(strlocaltime,0,sizeof(strlocaltime));localtime(strlocaltime,"yyyymmddhh24miss");char strxmlfilename[301];memset(strxmlfilename,0,sizeof(strxmlfilename));snprintf(strxmlfilename,300,"%s/tbspaceinfo_%s_%s.xml",argv[3],strlocaltime,argv[1]);cfile xmlfile;if (xmlfile.openforrename(strxmlfilename,"w ") == false ){logfile.write("xmlfile.openforrename(%s) failed.\n",strxmlfilename); return -1;}xmlfile.fprintf("\n");while (true){memset(&sttbspaceinfo,0,sizeof(sttbspaceinfo));if (stmt.next() != 0) break;xmlfile.fprintf(\"%s"\"%s"\"%s"\"%0.02f"\"%0.02f"\"%0.02f"\"%0.02f\n",argv[1],strlocaltime,sttbspaceinfo.tablespace,sttbspaceinfo.total,sttbspaceinfo.used,sttbspaceinfo.available,sttbspaceinfo.usep);}xmlfile.fprintf("\n");xmlfile.closeandrename();logfile.write("生成文件%s.\n",strxmlfilename);exit(0); } void exit(int sig) {if (sig > 0) signal(sig,sig_ign);logfile.write("catching the signal(%d).\n",sig);logfile.write("tbspaceinfo exit.\n");exit(0); }


以下是收集oracle会话信息,我们用客户端通oracle的监听连上oracle数据库,oracle数据库会启动一个进程向会话提供服务,是多进程的服务端,每增加一连接(进程),需要消耗系统资源(内存,socket连接)。如下一共有6个客户端连上,sql>exit退出一个就剩下5个。ps -ef |grep oracle可查看oracle数据库系统进程。


如下两个sql是查询数据字典查到进程和上面一样,第一个sql是查全部包括系统进程。如下表有一个osuser表示连上来的客户端操作系统名字,还可以通过sql_id查到连上来的客户端sql语句操作。还可以指定oracle用户的会话数。

// dbsessioninfo.cpp #include "qxmon.h" void exit(int sig); struct st_dbsessioninfo stdbsessioninfo; clogfile logfile; connection conndst; cprogramactive programactive;int main(int argc,char *argv[]) {if (argc != 5){printf("\n");printf("using:./dbsessioninfo hostname logfilename outputpath username/password@tnsnames\n");printf("example:/htidc/htidc/bin/procctl 300 /htidc/qxmon/bin/dbsessioninfo 10.153.98.13 /log/qxmon/dbsessioninfo_10.153.98.13.log /qxdata/qxmon/qxmonclient szidc/pwdidc@szqx_10.153.98.13\n\n");printf("此程序连接远程数据库,把远程数据库会话信息写入xml文件。\n");printf("hostname是本服务器的主机名,为了方便识别,也可以用ip。\n");printf("logfilename是本程序的日志文件名。\n");printf("outputpath是输出的xml文件存放的目录。\n");printf("username/password@tnsnames为待监控的远程数据库的用户名/密码@连接名。\n");printf("此程序运行在监控平台的服务器上。\n\n\n");return -1;}closeioandsignal(); signal(sigint,exit); signal(sigterm,exit);if (logfile.open(argv[2],"a ") == false){printf("logfile.open(%s) failed.\n",argv[2]); return -1;}programactive.setprograminfo(&logfile,"alarmserver",300);if (conndst.connecttodb(argv[4])!=0){logfile.write("connmon.connecttodb(%s) failed.\n",argv[4]); exit(-1);}sqlstatement stmt;stmt.connect(&conndst);stmt.prepare("select count(*) from v$session where type!='background' and status!='killed'"); stmt.bindout(1,&stdbsessioninfo.total);if (stmt.execute() != 0){logfile.write("select v$session failed.\n%s\n",stmt.cda.message); exit(-1);}char strlocaltime[21];memset(strlocaltime,0,sizeof(strlocaltime));localtime(strlocaltime,"yyyymmddhh24miss");char strxmlfilename[301];memset(strxmlfilename,0,sizeof(strxmlfilename));snprintf(strxmlfilename,300,"%s/dbsessioninfo_%s_%s.xml",argv[3],strlocaltime,argv[1]);cfile xmlfile;if (xmlfile.openforrename(strxmlfilename,"w ") == false ){logfile.write("xmlfile.openforrename(%s) failed.\n",strxmlfilename); exit(-1);}xmlfile.fprintf("\n");while (true){memset(&stdbsessioninfo,0,sizeof(stdbsessioninfo));if (stmt.next() != 0) break;xmlfile.fprintf(\"%s"\"%s"\"%d\n",argv[1],strlocaltime,stdbsessioninfo.total);}xmlfile.fprintf("\n");xmlfile.closeandrename();logfile.write("生成文件%s.\n",strxmlfilename);exit(0); }void exit(int sig) {if (sig > 0) signal(sig,sig_ign);logfile.write("catching the signal(%d).\n",sig);logfile.write("dbsessioninfo exit.\n");exit(0); }

oracle里有一个系统参数processes(最大允许的会话总数)。_ooci.h中connection类连接上一个后进程及processes 1,sqlstatement类实例化并指定数据库连接也会消耗一种资源即open_cursors参数。oracle有缓存机制存sql,下次用到则不进行语法分析,所以oracle需大内存,存的越多,缓存命中率越高。

总结

以上是ag凯发k8国际为你收集整理的【c/c 13】天气app:数据挖掘/http协议/非结构化数据存储(filetoblob.cpp),数据管理/监控告警(hsmtable.cpp,tbspaceinfo.cpp)的全部内容,希望文章能够帮你解决所遇到的问题。

如果觉得ag凯发k8国际网站内容还不错,欢迎将ag凯发k8国际推荐给好友。

  • 上一篇:
  • 下一篇:
网站地图