lower and lower performance when restore logical log

Hi every one: I have a question about restore logical log. I have two identical instance on two box,one box is HP NetServer LX PRO with SCO UNIX 5.0.4 and IDS 7.30 ,another box is IBM Netfinity 7000 TM0 with SCO UNIX 5.0.4 and IDS 7.30. HP box is our production machine, IBM box is our backup machine. I backup database with following steps: 1. do level 0 archive backup to tape on the HP box; 2. using 'ontape -c' to backup logical log to disk on the HP box; 3. using 'ontape -r' to restore level 0 archive on the IBM box, and when prompt 'Do you want restore logical log ? (Y/N)' I wait here; 4. stop logical log backup on the HP box with delete key; 5. on the IBM box press Y to restore logical log through network; 6. when step 5 prompt 'Do you want restore another logical log (Y/N) ' wait here, and move logical log file to a file naming logical.xxx-xxx on the HP box ,prepare another file and using 'ontape -c' to continue backup logical log on the HP box; 7. repeating step4-6 to gain a synchronization of two box. my LTAPESIZE is 10 MB,and block size is 1024K on the two box, at beginning it use 25 seconds to complete one log restore, but restore times is longer and longer while number of log increasing , it is about 40 minutes to restore one log after restoring 250 logs. Why online performance getting lower and lower when continue restore logical log on the IBM box ? How to gain best performance when continue restore logical log? IBM box is not run another programs except restore logical log.

In article <77qhgj$9jq$1@news.xmission.com>, liuaihui@163.net <liuaihui@163.net> writes > >Hi every one: > I have a question about restore logical log. I have two identical >instance on two box,one box is HP NetServer LX PRO with SCO UNIX 5.0.4 >and IDS 7.30 ,another box is IBM Netfinity 7000 TM0 with SCO UNIX 5.0.4 >and IDS 7.30. HP box is our production machine, IBM box is our backup >machine. I backup database with following steps: >1. do level 0 archive backup to tape on the HP box; >2. using 'ontape -c' to backup logical log to disk on the HP box; >3. using 'ontape -r' to restore level 0 archive on the IBM box, and when >prompt 'Do you want restore logical log ? (Y/N)' I wait here; >4. stop logical log backup on the HP box with delete key; >5. on the IBM box press Y to restore logical log through network; >6. when step 5 prompt 'Do you want restore another logical log (Y/N) ' >wait here, and move logical log file to a file naming logical.xxx-xxx on >the HP box ,prepare another file and using 'ontape -c' to continue >backup logical log on the HP box; >7. repeating step4-6 to gain a synchronization of two box. > >my LTAPESIZE is 10 MB,and block size is 1024K on the two box, at >beginning it use 25 seconds to complete one log restore, but restore >times is longer and longer while number of log increasing , >it is about 40 minutes to restore one log after restoring 250 logs. Why >online performance getting lower and lower when continue restore logical >log on the IBM box ? How to gain best performance when continue restore >logical log? > IBM box is not run another programs except restore logical log. Don't know, run onstat command to see what the IBM box is doing. Also talk to Informix Technical Support. -- David Williams