Chinaunix
标题:
还AIX上面的Process问题
[打印本页]
作者:
chao_ping
时间:
2001-11-29 11:46
标题:
还AIX上面的Process问题
原来那个帖子太长了,另起一个线程:\r\nAIX 4.3.3\r\nOracle 8.1.7.2\r\n症状:\r\nOralce用户有时候不能启动多个进程。\r\n比如:\r\n\r\nSQL> startup nomount\r\nORACLE instance started.\r\n\r\nTotal System Global Area 888225684 bytes\r\nFixed Size 73620 bytes\r\nVariable Size 438902784 bytes\r\nDatabase Buffers 448192512 bytes\r\nRedo Buffers 1056768 bytes\r\nSQL> show parameter job\r\n\r\nNAME TYPE VALUE\r\n------------------------------------ ------- ------------------------------\r\njob_queue_interval integer 60\r\njob_queue_processes integer 1\r\nSQL> alter database mount;\r\n\r\n
<p><b><font size=\"4\" color=\"#FF00FF\">0403-030 The fork function failed. Too many processes already exist.<br> \r\n</font></b></p>
\r\nDatabase altered.\r\n\r\nSQL> alter database open;\r\n0403-030 The fork function failed. Too many processes already exist.\r\n\r\nDatabase altered.\r\n大家需要那些参数来帮我诊断?\r\n\r\n
作者:
chao_ping
时间:
2001-11-29 12:22
补充:\r\n1.当前用户只有一个连接,所以应该不是License问题.\r\napp_svc
raids> who\r\nroot lft0 Nov 25 18:46\r\noraids pts/0 Nov 29 12:20 (s80_svc)\r\nroot pts/1 Nov 25 18:48 (:0.0)\r\nroot pts/3 Nov 27 13:22 (:0.0)\r\n2.当前用户全部启动的进程数目不到13个\r\napp_svc
raids> ps -ef|grep oraids\r\n oraids 143558 1 0 11:50:27 - 0:00 /oracle/IDS/817_32/bin/tnslsnr LISTENER -inherit\r\n oraids 144848 1 0 11:25:22 - 0:00 ora_arc0_IDS\r\n oraids 145150 1 0 11:25:22 - 0:00 ora_snp0_IDS\r\n oraids 153166 1 0 11:25:22 - 0:00 ora_pmon_IDS\r\n oraids 154678 1 0 11:25:22 - 0:00 ora_lgwr_IDS\r\n oraids 157010 295732 4 12:21:06 pts/0 0:00 ps -ef\r\n oraids 158996 1 0 11:25:22 - 0:00 ora_ckpt_IDS\r\n oraids 289564 1 0 11:25:22 - 0:00 ora_smon_IDS\r\n oraids 292088 295732 0 12:21:06 pts/0 0:00 grep oraids\r\n oraids 293422 1 0 11:25:22 - 0:00 ora_reco_IDS\r\n oraids 293994 1 0 11:25:22 - 0:00 ora_dbw0_IDS\r\n oraids 295732 294830 0 12:20:23 pts/0 0:00 -ksh\r\n3. 有了这么几个进程以后,只能够再启动一个进程:\r\nsqlplus internal\r\n然胡想要Host都不行了.\r\napp_svc
raids> sqlplus system/manage\r\n\r\nSQL*Plus: Release 8.1.7.0.0 - Production on Thu Nov 29 12:21:44 2001\r\n\r\n(c) Copyright 2000 Oracle Corporation. All rights reserved.\r\n\r\n\r\nConnected to:\r\nOracle8i Enterprise Edition Release 8.1.7.2.0 - Production\r\nWith the Partitioning option\r\nJServer Release 8.1.7.2.0 - Production\r\n\r\nSQL> !\r\n/usr/bin/ksh: Resource temporarily unavailable\r\n我exit出来才可以.\r\n再启动几个进程就不行,比如运行三次sh,就报那个错误.\r\napp_svc
raids> sh\r\napp_svc
raids> sh\r\napp_svc
raids> sh .过很久,出现\r\nsh: 0403-030 The fork function failed. Too many processes already exist\r\napp_svc
raids>\r\n这个时候从客户端连接,出现:\r\nC:\\>sqlplus system/manage@ids\r\n\r\nSQL*Plus: Release 8.1.7.0.0 - Production on Thu Nov 29 12:27:56 2001\r\n\r\n(c) Copyright 2000 Oracle Corporation. All rights reserved.\r\n\r\nERROR:\r\nORA-12500: TNS:listener failed to start a dedicated server process\r\n\r\n\r\nEnter user-name:
作者:
chao_ping
时间:
2001-11-29 16:04
标题:
补充:
Oracle用户的ulimits情况:\r\n\r\napp_svc
raids> ulimit -a\r\ntime(seconds) unlimited\r\nfile(blocks) unlimited\r\ndata(kbytes) unlimited\r\nstack(kbytes) unlimited\r\nmemory(kbytes) 32768\r\ncoredump(blocks) 2097151\r\nnofiles(descriptors) 2000\r\napp_svc
raids>\r\n\r\n系统属性:\r\napp_svc
raids> lsattr -El sys0\r\nkeylock normal State of system keylock at boot time False\r\nmaxbuf 200 Maximum number of pages in block I/O BUFFER CACHE True\r\nmaxmbuf 0 Maximum Kbytes of real memory allowed for MBUFS True\r\nmaxuproc 512 Maximum number of PROCESSES allowed per user True\r\nautorestart false Automatically REBOOT system after a crash True\r\niostat true Continuously maintain DISK I/O history True\r\nrealmem 4194304 Amount of usable physical memory in Kbytes False\r\nconslogin enable System Console Login False\r\nfwversion IBM,M2P01072 Firmware version and revision levels False\r\nmaxpout 0 HIGH water mark for pending write I/Os per file True\r\nminpout 0 LOW water mark for pending write I/Os per file True\r\nfullcore false Enable full CORE dump True\r\npre430core false Use pre-430 style CORE dump True\r\nrtasversion 1 Open Firmware RTAS version False\r\nmodelname IBM,7026-H80 Machine name False\r\nsystemid IBM,01108C5AF Hardware system identifier False\r\nboottype disk N/A False\r\nSW_dist_intr false Enable SW distribution of interrupts True\r\ncpuguard disable CPU Guard True
作者:
slls
时间:
2001-11-29 16:08
你改成512后reboot了吧?\r\n只有oracle用户这样吗?
作者:
chao_ping
时间:
2001-11-29 16:15
reboot过几次了
\r\n\r\n好像只有Oracle用户这样,别的没有什么用户啊,也没法测试。\r\n另外的sap的用户idsadm倒是正常的。\r\n我现在地暂时办法就是用sap用户启动oracle,启动sap。
作者:
slls
时间:
2001-11-29 16:36
真见鬼了!\r\n你先把oracle下来,再不停地sh,sh,sh,sh...看看问题是否出现,如果没事,那问题应该在oracle吧?
作者:
slls
时间:
2001-11-29 16:37
或者把.profile中与oracle启动没有直接关系的语句先注掉?
作者:
chao_ping
时间:
2001-11-29 19:14
还是一样。\r\n我关闭Oracle,然后su oraids,这样的话,profile文件就是不起作用的。\r\n然后我sh 12次以后,\r\nroot@app_svc#su oraids\r\n$ sh\r\n$ sh\r\n$ sh\r\n$ sh\r\n$ sh\r\n$ sh\r\n$ sh\r\n$ sh\r\n$ sh\r\n$ sh\r\n$ sh\r\n$ sh\r\n\r\nsh: 0403-030 The fork function failed. Too many processes already exist.\r\n$ ^C$\r\n再连ls夜运行不了了。只能ctrl-C中止。\r\n然后我su 到root,\r\n$ exit\r\n$ su\r\nroot\'s Password:\r\n# who\r\nroot lft0 Nov 25 18:46\r\nroot pts/0 Nov 29 15:02 (192.168.8.103)\r\nroot pts/1 Nov 25 18:48 (:0.0)\r\nroot pts/2 Nov 29 16:13 (s80_svc)\r\nroot pts/3 Nov 27 13:22 (:0.0)\r\n# ps -ef|grep ora\r\n oraids 144852 293426 0 19:15:49 pts/0 0:00 sh\r\n oraids 144898 154428 0 19:15:45 pts/0 0:00 sh\r\n oraids 148132 289568 0 19:15:48 pts/0 0:00 sh\r\n oraids 150974 158834 0 19:15:42 pts/0 0:00 ksh\r\n oraids 154428 293278 0 19:15:45 pts/0 0:00 sh\r\n idsadm 157964 1 0 16:23:13 pts/0 0:00 /oracle/IDS/817_32/bin/t\r\n oraids 159010 144898 0 19:15:46 pts/0 0:00 sh\r\n oraids 160080 150974 0 19:15:43 pts/0 0:00 sh\r\n oraids 289568 159010 0 19:15:47 pts/0 0:00 sh\r\n root 290334 165374 0 19:18:16 pts/0 0:00 grep ora\r\n oraids 293278 301410 0 19:15:44 pts/0 0:00 sh\r\n oraids 293426 148132 0 19:15:48 pts/0 0:00 sh\r\n oraids 301410 160080 0 19:15:44 pts/0 0:00 sh\r\n应该是何Oracle没有关系。\r\n所以我贴到AIX版面来了。
作者:
overtime
时间:
2001-11-29 23:04
Aix每个用户都有process数限制,可以用smit修改,不知是不是这个原因.
作者:
larryh
时间:
2001-11-30 00:25
有可能是BUG,你的Maintenance Level打到多少?
作者:
slls
时间:
2001-11-30 11:05
看来没有别的可以解释这个问题了。要么打补丁要么打800吧。
作者:
chao_ping
时间:
2001-11-30 16:00
AIX的补丁从哪里下载?
欢迎光临 Chinaunix (http://bbs.chinaunix.net/)
Powered by Discuz! X3.2