EXP/IMP massive data

EXP/IMP massive data

From: Macro Beth Oracle exp/IMP are two tools frequently used by many users. they are often used for database logical backup, database restructuring, data transfer, and other work. first, exp unload the data to the file system to generate. DMP file, and then the data is loaded into the database by IMP when necessary. for small and medium databases, the DMP file generated by the full database exp may be less than 2 GB, but for a slightly large database, the data generated by exp may be dozens to hundreds of GB. currently, most operating systems are 32-bit, and the maximum file size allowed by the file system is 2 GB. obviously, the file system cannot store exp data. this is one of the problems. another problem is that as the database grows, exp takes longer and longer to implement. this article discusses the corresponding countermeasures for the above two problems.   Taking Unix as an example, let's first look at how to surpass the limit of 2 GB. this requires the use of common UNIX technologies, such as pipelines (Named Pipe), data copy tool dd, and data compression (compress ). these technologies are discussed below. · PipelinesIs a pseudo file. it exists in the memory and is used for fast I/O operations. the buffer zone of the MPs queue adopts the first-in-first-out mechanism, that is, the write pipeline process writes the buffer header and the read pipeline process reads the tail of the MPs queue. the pipeline creation command is "mknod filename P ". Dd allows us to copy data from one device to another. · Compress is a UNIX Data Compression tool. Before implementing exp, we can check the size of the generated DMP file. The following steps can be done, 1. Generate MPs queue: $ Mknod/tmp/mypipe P 2. The exp result is output to the MPs queue. $ Exp file =/tmp/mypipe <... Other exp Options> & 3. Read the MPs queue and output the result to the MPs queue to view the data volume: $ Dd If =/tmp/mypipe of =/dev/null The size of the exp (. dmp) file returned, in the unit of 512 bytes ). Now we can start to discuss specific methods, 1. File compression 2. Directly exp to tape 3. Comprehensive Methods First, let's take a look at the file compression method. With File compression, we can try to avoid the DMP file size exceeding the limit of 2 GB. We use the Unix pipeline technology, the specific steps are as follows, 1. Start the compress process to read data from the pipeline and output the data to the disk file. $ Mknod/tmp/exp_pipe P $ Compress </tmp/exp_pipe> export. dmp. Z & 2. exp to MPs queue. $ Exp file =/tmp/exp_pipe... & IMP is similar. Let's take a look at the method to directly import exp to the tape. $ Exp file =/dev/RMT/0 m volsize = 4G $ Imp file =/dev/RMT/0 m volsize = 4G Finally, let's take a look at the exp method on the bare device, $ Exp file =/dev/rdsk/c0t3d0s0 volsize = 4G IMP is similar. The above method can avoid the limit of 2 GB.   The following describes how to shorten the exp Cable Time. at present, most users use the exp method to tape because there is not enough hard disk space. most of the time spent in the entire exp process is written on tape. through the discussion of the UNIX tool, we can first export the data to the MPs queue, compress the MPs queue data, and then output the data to the tape. in this way, the amount of data written to the tape can be greatly reduced, thus shortening the time to write the tape to a considerable extent. in addition, from an internal perspective of Oracle, starting from Version 7.3, Oracle allows users to perform direct path export, that is, skipping Oracle $ Make-F $ ORACLE_HOME/rdbms/lib/oracle. mk expst $ Make-F $ ORACLE_HOME/rdbms/lib/oracle. mk impst The executable files expst (exp single task) and impst generated in the preceding steps are used exactly the same as exp/imp. using them can save up to 30% of the time than using exp/imp. however, you must note that when using expst/impst, you must ensure that no other user is using the Oracle database, otherwise, the database may have unpredictable consequences or even destruction. $ Make-F $ ORACLE_HOME/rdbms/lib/oracle. mk expst $ Make-F $ ORACLE_HOME/rdbms/lib/oracle. mk impst The executable files expst (exp single task) and impst generated in the preceding steps are used exactly the same as exp/imp. using them can save up to 30% of the time than using exp/imp. however, you must note that when using expst/impst, you must ensure that no other user is using the Oracle database, otherwise, the database may have unpredictable consequences or even destruction. Before implementing exp, we can check the size of the generated DMP file. The following steps can be done, 1. Generate MPs queue: $ Mknod/tmp/mypipe P 2. The exp result is output to the MPs queue. $ Exp file =/tmp/mypipe <... Other exp Options> & 3. Read the MPs queue and output the result to the MPs queue to view the data volume: $ Dd If =/tmp/mypipe of =/dev/null The size of the exp (. dmp) file returned, in the unit of 512 bytes ). Now we can start to discuss specific methods, 1. File compression 2. Directly exp to tape 3. Comprehensive Methods First, let's take a look at the file compression method. With File compression, we can try to avoid the DMP file size exceeding the limit of 2 GB. We use the Unix pipeline technology, the specific steps are as follows, 1. Start the compress process to read data from the pipeline and output the data to the disk file. $ Mknod/tmp/exp_pipe P $ Compress </tmp/exp_pipe> export. dmp. Z & 2. exp to MPs queue. $ Exp file =/tmp/exp_pipe... & IMP is similar. Let's take a look at the method to directly import exp to the tape. $ Exp file =/dev/RMT/0 m volsize = 4G $ Imp file =/dev/RMT/0 m volsize = 4G Finally, let's take a look at the exp method on the bare device, $ Exp file =/dev/rdsk/c0t3d0s0 volsize = 4G IMP is similar. The above method can avoid the limit of 2 GB.   The following describes how to shorten the exp Cable Time. at present, most users use the exp method to tape because there is not enough hard disk space. most of the time spent in the entire exp process is written on tape. through the discussion of the UNIX tool, we can first export the data to the MPs queue, compress the MPs queue data, and then output the data to the tape. in this way, the amount of data written to the tape can be greatly reduced, thus shortening the time to write the tape to a considerable extent. in addition, from an internal perspective of Oracle, starting from Version 7.3, Oracle allows users to perform direct path export, that is, skipping Oracle $ Make-F $ ORACLE_HOME/rdbms/lib/oracle. mk expst $ Make-F $ ORACLE_HOME/rdbms/lib/oracle. mk impst The executable files expst (exp single task) and impst generated in the preceding steps are used exactly the same as exp/imp. using them can save up to 30% of the time than using exp/imp. however, you must note that when using expst/impst, you must ensure that no other user is using the Oracle database, otherwise, the database may have unpredictable consequences or even destruction. $ Make-F $ ORACLE_HOME/rdbms/lib/oracle. mk expst $ Make-F $ ORACLE_HOME/rdbms/lib/oracle. mk impst The executable files expst (exp single task) and impst generated in the preceding steps are used exactly the same as exp/imp. using them can save up to 30% of the time than using exp/imp. however, you must note that when using expst/impst, you must ensure that no other user is using the Oracle database, otherwise, the database may have unpredictable consequences or even destruction. · Compress is a UNIX Data Compression tool. Before implementing exp, we can check the size of the generated DMP file. The following steps can be done, 1. Generate MPs queue: $ Mknod/tmp/mypipe P 2. The exp result is output to the MPs queue. $ Exp file =/tmp/mypipe <... Other exp Options> & 3. Read the MPs queue and output the result to the MPs queue to view the data volume: $ Dd If =/tmp/mypipe of =/dev/null The size of the exp (. dmp) file returned, in the unit of 512 bytes ). Now we can start to discuss specific methods, 1. File compression 2. Directly exp to tape 3. Comprehensive Methods First, let's take a look at the file compression method. With File compression, we can try to avoid the DMP file size exceeding the limit of 2 GB. We use the Unix pipeline technology, the specific steps are as follows, 1. Start the compress process to read data from the pipeline and output the data to the disk file. $ Mknod/tmp/exp_pipe P $ Compress </tmp/exp_pipe> export. dmp. Z & 2. exp to MPs queue. $ Exp file =/tmp/exp_pipe... & IMP is similar. Let's take a look at the method to directly import exp to the tape. $ Exp file =/dev/RMT/0 m volsize = 4G $ Imp file =/dev/RMT/0 m volsize = 4G Finally, let's take a look at the exp method on the bare device, $ Exp file =/dev/rdsk/c0t3d0s0 volsize = 4G IMP is similar. The above method can avoid the limit of 2 GB.   The following describes how to shorten the exp Cable Time. at present, most users use the exp method to tape because there is not enough hard disk space. most of the time spent in the entire exp process is written on tape. through the discussion of the UNIX tool, we can first export the data to the MPs queue, compress the MPs queue data, and then output the data to the tape. in this way, the amount of data written to the tape can be greatly reduced, thus shortening the time to write the tape to a considerable extent. in addition, from an internal perspective of Oracle, starting from Version 7.3, Oracle allows users to perform direct path export, that is, skipping Oracle $ Make-F $ ORACLE_HOME/rdbms/lib/oracle. mk expst $ Make-F $ ORACLE_HOME/rdbms/lib/oracle. mk impst The executable files expst (exp single task) and impst generated in the preceding steps are used exactly the same as exp/imp. using them can save up to 30% of the time than using exp/imp. however, you must note that when using expst/impst, you must ensure that no other user is using the Oracle database, otherwise, the database may have unpredictable consequences or even destruction. $ Make-F $ ORACLE_HOME/rdbms/lib/oracle. mk expst $ Make-F $ ORACLE_HOME/rdbms/lib/oracle. mk impst The executable files expst (exp single task) and impst generated in the preceding steps are used exactly the same as exp/imp. using them can save up to 30% of the time than using exp/imp. however, you must note that when using expst/impst, you must ensure that no other user is using the Oracle database, otherwise, the database may have unpredictable consequences or even destruction. Before implementing exp, we can check the size of the generated DMP file. The following steps can be done, 1. Generate MPs queue: $ Mknod/tmp/mypipe P 2. The exp result is output to the MPs queue. $ Exp file =/tmp/mypipe <... Other exp Options> & 3. Read the MPs queue and output the result to the MPs queue to view the data volume: $ Dd If =/tmp/mypipe of =/dev/null The size of the exp (. dmp) file returned, in the unit of 512 bytes ). Now we can start to discuss specific methods, 1. File compression 2. Directly exp to tape 3. Comprehensive Methods First, let's take a look at the file compression method. With File compression, we can try to avoid the DMP file size exceeding the limit of 2 GB. We use the Unix pipeline technology, the specific steps are as follows, 1. Start the compress process to read data from the pipeline and output the data to the disk file. $ Mknod/tmp/exp_pipe P $ Compress </tmp/exp_pipe> export. dmp. Z & 2. exp to MPs queue. $ Exp file =/tmp/exp_pipe... & IMP is similar. Let's take a look at the method to directly import exp to the tape. $ Exp file =/dev/RMT/0 m volsize = 4G $ Imp file =/dev/RMT/0 m volsize = 4G Finally, let's take a look at the exp method on the bare device, $ Exp file =/dev/rdsk/c0t3d0s0 volsize = 4G IMP is similar. The above method can avoid the limit of 2 GB.   The following describes how to shorten the exp Cable Time. at present, most users use the exp method to tape because there is not enough hard disk space. most of the time spent in the entire exp process is written on tape. through the discussion of the UNIX tool, we can first export the data to the MPs queue, compress the MPs queue data, and then output the data to the tape. in this way, the amount of data written to the tape can be greatly reduced, thus shortening the time to write the tape to a considerable extent. in addition, from an internal perspective of Oracle, starting from Version 7.3, Oracle allows users to perform direct path export, that is, skipping Oracle $ Make-F $ ORACLE_HOME/rdbms/lib/oracle. mk expst $ Make-F $ ORACLE_HOME/rdbms/lib/oracle. mk impst The executable files expst (exp single task) and impst generated in the preceding steps are used exactly the same as exp/imp. using them can save up to 30% of the time than using exp/imp. however, you must note that when using expst/impst, you must ensure that no other user is using the Oracle database, otherwise, the database may have unpredictable consequences or even destruction. $ Make-F $ ORACLE_HOME/rdbms/lib/oracle. mk expst $ Make-F $ ORACLE_HOME/rdbms/lib/oracle. mk impst The executable files expst (exp single task) and impst generated in the preceding steps are used exactly the same as exp/imp. using them can save up to 30% of the time than using exp/imp. however, you must note that when using expst/impst, you must ensure that no other user is using the Oracle database, otherwise, the database may have unpredictable consequences or even destruction. Dd allows us to copy data from one device to another. · Compress is a UNIX Data Compression tool. Before implementing exp, we can check the size of the generated DMP file. The following steps can be done, 1. Generate MPs queue: $ Mknod/tmp/mypipe P 2. The exp result is output to the MPs queue. $ Exp file =/tmp/mypipe <... Other exp Options> & 3. Read the MPs queue and output the result to the MPs queue to view the data volume: $ Dd If =/tmp/mypipe of =/dev/null The size of the exp (. dmp) file returned, in the unit of 512 bytes ). Now we can start to discuss specific methods, 1. File compression 2. Directly exp to tape 3. Comprehensive Methods First, let's take a look at the file compression method. With File compression, we can try to avoid the DMP file size exceeding the limit of 2 GB. We use the Unix pipeline technology, the specific steps are as follows, 1. Start the compress process to read data from the pipeline and output the data to the disk file. $ Mknod/tmp/exp_pipe P $ Compress </tmp/exp_pipe> export. dmp. Z & 2. exp to MPs queue. $ Exp file =/tmp/exp_pipe... & IMP is similar. Let's take a look at the method to directly import exp to the tape. $ Exp file =/dev/RMT/0 m volsize = 4G $ Imp file =/dev/RMT/0 m volsize = 4G Finally, let's take a look at the exp method on the bare device, $ Exp file =/dev/rdsk/c0t3d0s0 volsize = 4G IMP is similar. The above method can avoid the limit of 2 GB.   The following describes how to shorten the exp Cable Time. at present, most users use the exp method to tape because there is not enough hard disk space. most of the time spent in the entire exp process is written on tape. through the discussion of the UNIX tool, we can first export the data to the MPs queue, compress the MPs queue data, and then output the data to the tape. in this way, the amount of data written to the tape can be greatly reduced, thus shortening the time to write the tape to a considerable extent. in addition, from an internal perspective of Oracle, starting from Version 7.3, Oracle allows users to perform direct path export, that is, skipping Oracle $ Make-F $ ORACLE_HOME/rdbms/lib/oracle. mk expst $ Make-F $ ORACLE_HOME/rdbms/lib/oracle. mk impst The executable files expst (exp single task) and impst generated in the preceding steps are used exactly the same as exp/imp. using them can save up to 30% of the time than using exp/imp. however, you must note that when using expst/impst, you must ensure that no other user is using the Oracle database, otherwise, the database may have unpredictable consequences or even destruction. $ Make-F $ ORACLE_HOME/rdbms/lib/oracle. mk expst $ Make-F $ ORACLE_HOME/rdbms/lib/oracle. mk impst The executable files expst (exp single task) and impst generated in the preceding steps are used exactly the same as exp/imp. using them can save up to 30% of the time than using exp/imp. however, you must note that when using expst/impst, you must ensure that no other user is using the Oracle database, otherwise, the database may have unpredictable consequences or even destruction. Before implementing exp, we can check the size of the generated DMP file. The following steps can be done, 1. Generate MPs queue: $ Mknod/tmp/mypipe P 2. The exp result is output to the MPs queue. $ Exp file =/tmp/mypipe <... Other exp Options> & 3. Read the MPs queue and output the result to the MPs queue to view the data volume: $ Dd If =/tmp/mypipe of =/dev/null The size of the exp (. dmp) file returned, in the unit of 512 bytes ). Now we can start to discuss specific methods, 1. File compression 2. Directly exp to tape 3. Comprehensive Methods First, let's take a look at the file compression method. With File compression, we can try to avoid the DMP file size exceeding the limit of 2 GB. We use the Unix pipeline technology, the specific steps are as follows, 1. Start the compress process to read data from the pipeline and output the data to the disk file. $ Mknod/tmp/exp_pipe P $ Compress </tmp/exp_pipe> export. dmp. Z & 2. exp to MPs queue. $ Exp file =/tmp/exp_pipe... & IMP is similar. Let's take a look at the method to directly import exp to the tape. $ Exp file =/dev/RMT/0 m volsize = 4G $ Imp file =/dev/RMT/0 m volsize = 4G Finally, let's take a look at the exp method on the bare device, $ Exp file =/dev/rdsk/c0t3d0s0 volsize = 4G IMP is similar. The above method can avoid the limit of 2 GB.   The following describes how to shorten the exp Cable Time. at present, most users use the exp method to tape because there is not enough hard disk space. most of the time spent in the entire exp process is written on tape. through the discussion of the UNIX tool, we can first export the data to the MPs queue, compress the MPs queue data, and then output the data to the tape. in this way, the amount of data written to the tape can be greatly reduced, thus shortening the time to write the tape to a considerable extent. in addition, from an internal perspective of Oracle, starting from Version 7.3, Oracle allows users to perform direct path export, that is, skipping Oracle $ Make-F $ ORACLE_HOME/rdbms/lib/oracle. mk expst $ Make-F $ ORACLE_HOME/rdbms/lib/oracle. mk impst The executable files expst (exp single task) and impst generated in the preceding steps are used exactly the same as exp/imp. using them can save up to 30% of the time than using exp/imp. however, you must note that when using expst/impst, you must ensure that no other user is using the Oracle database, otherwise, the database may have unpredictable consequences or even destruction. $ Make-F $ ORACLE_HOME/rdbms/lib/oracle. mk expst $ Make-F $ ORACLE_HOME/rdbms/lib/oracle. mk impst The executable files expst (exp single task) and impst generated in the preceding steps are used exactly the same as exp/imp. using them can save up to 30% of the time than using exp/imp. however, you must note that when using expst/impst, you must ensure that no other user is using the Oracle database, otherwise, the database may have unpredictable consequences or even destruction. · Compress is a UNIX Data Compression tool. Before implementing exp, we can check the size of the generated DMP file. The following steps can be done, 1. Generate MPs queue: $ Mknod/tmp/mypipe P 2. The exp result is output to the MPs queue. $ Exp file =/tmp/mypipe <... Other exp Options> & 3. Read the MPs queue and output the result to the MPs queue to view the data volume: $ Dd If =/tmp/mypipe of =/dev/null The size of the exp (. dmp) file returned, in the unit of 512 bytes ). Now we can start to discuss specific methods, 1. File compression 2. Directly exp to tape 3. Comprehensive Methods First, let's take a look at the file compression method. With File compression, we can try to avoid the DMP file size exceeding the limit of 2 GB. We use the Unix pipeline technology, the specific steps are as follows, 1. Start the compress process to read data from the pipeline and output the data to the disk file. $ Mknod/tmp/exp_pipe P $ Compress </tmp/exp_pipe> export. dmp. Z & 2. exp to MPs queue. $ Exp file =/tmp/exp_pipe... & IMP is similar. Let's take a look at the method to directly import exp to the tape. $ Exp file =/dev/RMT/0 m volsize = 4G $ Imp file =/dev/RMT/0 m volsize = 4G Finally, let's take a look at the exp method on the bare device, $ Exp file =/dev/rdsk/c0t3d0s0 volsize = 4G IMP is similar. The above method can avoid the limit of 2 GB.   The following describes how to shorten the exp Cable Time. at present, most users use the exp method to tape because there is not enough hard disk space. most of the time spent in the entire exp process is written on tape. through the discussion of the UNIX tool, we can first export the data to the MPs queue, compress the MPs queue data, and then output the data to the tape. in this way, the amount of data written to the tape can be greatly reduced, thus shortening the time to write the tape to a considerable extent. in addition, from an internal perspective of Oracle, starting from Version 7.3, Oracle allows users to perform direct path export, that is, skipping Oracle $ Make-F $ ORACLE_HOME/rdbms/lib/oracle. mk expst $ Make-F $ ORACLE_HOME/rdbms/lib/oracle. mk impst The executable files expst (exp single task) and impst generated in the preceding steps are used exactly the same as exp/imp. using them can save up to 30% of the time than using exp/imp. however, you must note that when using expst/impst, you must ensure that no other user is using the Oracle database, otherwise, the database may have unpredictable consequences or even destruction. $ Make-F $ ORACLE_HOME/rdbms/lib/oracle. mk expst $ Make-F $ ORACLE_HOME/rdbms/lib/oracle. mk impst The executable files expst (exp single task) and impst generated in the preceding steps are used exactly the same as exp/imp. using them can save up to 30% of the time than using exp/imp. however, you must note that when using expst/impst, you must ensure that no other user is using the Oracle database, otherwise, the database may have unpredictable consequences or even destruction. Before implementing exp, we can check the size of the generated DMP file. The following steps can be done, 1. Generate MPs queue: $ Mknod/tmp/mypipe P 2. The exp result is output to the MPs queue. $ Exp file =/tmp/mypipe <... Other exp Options> & 3. Read the MPs queue and output the result to the MPs queue to view the data volume: $ Dd If =/tmp/mypipe of =/dev/null The size of the exp (. dmp) file returned, in the unit of 512 bytes ). Now we can start to discuss specific methods, 1. File compression 2. Directly exp to tape 3. Comprehensive Methods First, let's take a look at the file compression method. With File compression, we can try to avoid the DMP file size exceeding the limit of 2 GB. We use the Unix pipeline technology, the specific steps are as follows, 1. Start the compress process to read data from the pipeline and output the data to the disk file. $ Mknod/tmp/exp_pipe P $ Compress </tmp/exp_pipe> export. dmp. Z & 2. exp to MPs queue. $ Exp file =/tmp/exp_pipe... & IMP is similar. Let's take a look at the method to directly import exp to the tape. $ Exp file =/dev/RMT/0 m volsize = 4G $ Imp file =/dev/RMT/0 m volsize = 4G Finally, let's take a look at the exp method on the bare device, $ Exp file =/dev/rdsk/c0t3d0s0 volsize = 4G IMP is similar. The above method can avoid the limit of 2 GB.   The following describes how to shorten the exp Cable Time. at present, most users use the exp method to tape because there is not enough hard disk space. most of the time spent in the entire exp process is written on tape. through the discussion of the UNIX tool, we can first export the data to the MPs queue, compress the MPs queue data, and then output the data to the tape. in this way, the amount of data written to the tape can be greatly reduced, thus shortening the time to write the tape to a considerable extent. in addition, from an internal perspective of Oracle, starting from Version 7.3, Oracle allows users to perform direct path export, that is, skipping Oracle $ Make-F $ ORACLE_HOME/rdbms/lib/oracle. mk expst $ Make-F $ ORACLE_HOME/rdbms/lib/oracle. mk impst The executable files expst (exp single task) and impst generated in the preceding steps are used exactly the same as exp/imp. using them can save up to 30% of the time than using exp/imp. however, you must note that when using expst/impst, you must ensure that no other user is using the Oracle database, otherwise, the database may have unpredictable consequences or even destruction. $ Make-F $ ORACLE_HOME/rdbms/lib/oracle. mk expst $ Make-F $ ORACLE_HOME/rdbms/lib/oracle. mk impst The executable files expst (exp single task) and impst generated in the preceding steps are used exactly the same as exp/imp. using them can save up to 30% of the time than using exp/imp. however, you must note that when using expst/impst, you must ensure that no other user is using the Oracle database, otherwise, the database may have unpredictable consequences or even destruction.

[Last update: 10/15/2001 07:54:08]

About this site │ contact Webmaster │ comments and suggestions │ cooperation │ map │ advertisement service on this site

Copyright (c) 2000-2001 copyright to Oracle technology network

E-MAIL: oradb@126.com

Http://www.oradb.net/