Thanks for the help from Mr K S Braunsdorf .<br>This problem is caused by the opened pipe, which should have been closed, in the c++ program.<br><br><br>thanks<br>James Li<br><br><br><div class="gmail_quote">2010/11/11 jian li <span dir="ltr"><<a href="mailto:lijian06nju@gmail.com">lijian06nju@gmail.com</a>></span><br>
<blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;"><div><div></div><div class="h5">hi, all:<br> I need to back up about 50 files, the size of which won't exceed 5m, every 10~15 minutes to four remote machines.<br>
<br>
The back up command is written in a shell script file and was executed
by the scheduling program with system() function. The scheduling program
is implemented with c++.<br>
The command as follow:<br> <i>rsync -az /home/admin/service/* admin@10.249.49.101:/news_hot_data</i><br><br> <b> At first, this works just fine.</b> But after about one or two days, rsync will stop at some place and the whole backup process stuck.<br>
The following is the output with -vv option when backup stop:<br> <i>opening connection using: ssh -l admin 10.249.49.101 rsync --server -vvlogDtprze.isf . /news_hot_data</i><br><br>
I used strace to track the system call, and I found select() was
invoked again and again, never end until the program was killed by
ctrl+c. The following is the output:<br>
......<br> 18477 select(1027, [255 1024], [], NULL, NULL) = 1 (in [1024]) <div> <br> 18477 read(1024, "\36\0\0\0", 16384) = 4 <br> 18477 select(1027, [255 1024], [255], NULL, NULL) = 1 (out [255])<br>
18477 write(255, "]\306\304\2315\r\346\314\26]\2\275\350|X\305X\216\361\"\301}\t\34\213\357GPS\360\214\370"..., 48) = 48 <br> 18477 select(1027, [255 1024], [], NULL, NULL) = 1 (in [255])<br>
18477 read(255, "l\210\377V\20\270\0270\276@\363N\366\n!\311\211\312\206\216\25\3\1\323\375\370\24\0174lM\312"..., 8192) = 48<br> 18477 select(1027, [255 1024], [1025], NULL, NULL) = 1 (out [1025]) <br>
18477 write(1025, "\35\0\0\0\335\226\333L", 8) = 8<br> 18477 select(1027, [255 1024], [], NULL, NULL <unfinished ...> <br> 18476 <... select resumed> ) = 0 (Timeout)<br>
18476 select(1026, [], [], NULL, {60, 0}) = 0 (Timeout)<br> 18476 select(1026, [], [], NULL, {60, 0}) = 0 (Timeout) <br> 18476 select(1026, [], [], NULL, {60, 0}) = 0 (Timeout)<br>
18476 select(1026, [], [], NULL, {60, 0}) = 0 (Timeout) <br> 18476 select(1026, [], [], NULL, {60, 0}) = 0 (Timeout)<br> 18476 select(1026, [], [], NULL, {60, 0}) = 0 (Timeout) <br>
18476 select(1026, [], [], NULL, {60, 0}) = 0 (Timeout) <br> 18476 select(1026, [], [], NULL, {60, 0}) = 0 (Timeout) <br> 18476 select(1026, [], [], NULL, {60, 0}) = 0 (Timeout) <br>
18476 select(1026, [], [], NULL, {60, 0}) = 0 (Timeout) <br> 18476 select(1026, [], [], NULL, {60, 0}) = 0 (Timeout) <br> ......<br>
I read the linux manual about select(), know it was used to
wait something to be ready, but I don't know what exactly it is waiting,
and it never will ready.<br> <b>And what made me more confused is that when I execute the script in the terminal, it still works!</b><br>
<br> I have been confused for a few days and try to find out the reason by google, but failed.<br><br><br>thanks<br>James li</div>
</div></div></blockquote></div><br>