the db_verify slow gradually after system restart

[Former Member]

dear consultants:

After do the disk except error check , we found that the disk read is slow gradually after restart the system. but we have not the performance analyse before the disk except error check ,so we can't how to

lead the disk read slowly.

we found the disk read slowly by the db_verify job is slow gradually.

the db_verify command is "brbackup -u / -c force -m all -t online -w only_dbv"

now we don't know the databse reason or hardware lead to the db_verify slow.

could you so kind help me check it?

[Former Member]

Hi

To check if the disk read of the database hardware is slow or not , you can check

ST04 -> Detail analysis -> File system request , in this view , u can check the average read time of your datafile from OS level.

Also you can check the OS Log from ST06 --> Detailed analysis --> OS log.

These views can give the information whether the disk read times are higher.

Thanks

Rishi

[Former Member]

hello:

i want to find the the rate of the disk read , i want to compare .

and i don't know this is reason with the disk or the database

[stefan_koehler]

Hello Xuewen,

> i want to find the the rate of the disk read , i want to compare .

Yes that's fine but yo don't tell us your OS or database version?

If you are using AIX or Linux you can use nmon to monitor and track your disk I/O on OS level:

http://www.ibm.com/developerworks/aix/library/au-analyze_aix/

If you want the average wait time for I/O in the database you can use the AWR on Oracle 10g (sapnote #853576) or you can use STATSPACK for Oracle 9i (sapnote #717484).

Regards

Stefan

[Former Member]

oh yes,sorry for that.

our system is windows server 2003, and database is oracle102.

thank you for your kindly support, i will read the note carefully.

thank you again

best regards

wang

[Former Member]

Hi,

Please check out following steps:

1. Check windows task manager for CPU, Memory, PF usage and currently running applications. You can also stop applications if there is any which can run later.

2. Check ST06 for memory, CPU and disk with highest response time.

Also check the possibility of running verify processes in parallel. Any number of these verify processes can run in parallel, using parameter exec_parallel, option -e.

Hope this helps.

Manoj

Edited by: Manoj Chintawar on Mar 27, 2009 10:11 AM

[Former Member]

hello：

about the first

1. Check windows task manager for CPU, Memory, PF usage and currently running applications. You can also stop applications if there is any which can run later (of course non-sap).

i have check it ,the cpu and the memory are very natural as normal.

2. Check ST06 for memory, CPU and disk with highest response time.

could you tell me how to check it with the history?

thank you for your support

best regards

wang

[Former Member]

Hi,

Did you try ST06->Detailed Analysis menu->Compare days and other buttons there?

Manoj

[Former Member]

Date Usr Sys Idl Pages in/h Pages out/h

avg avg avg max avg max avg

20090330 54 11 37 11,751,415 7,294,764 625,927 463,689

20090329 6 5 89 19,303,440 5,172,478 57,429 51,052

20090328 4 5 93 15,745,216 6,043,112 57,410 57,134

20090327 52 10 38 11,105,000 7,445,815 508,025 401,639

20090326 54 11 35 12,216,027 6,900,755 810,504 563,203

20090325 47 10 44 11,213,059 7,052,218 479,928 377,111

20090324 71 11 18 457,531 319,927 664,318 566,415

20090323 74 12 15 5,924,609 4,389,900 870,460 692,559

20090322 25 9 66 15,144,538 8,214,990 56,176 39,043

20090321 23 7 71 17,190,030 7,791,974 57,329 50,367

20090320 68 12 20 7,606,063 5,189,594 521,834 356,468

20090319 56 10 34 7,778,120 5,381,566 385,264 289,993

20090318 49 7 44 434,529 286,035 733,312 408,583

20090317 48 11 42 1,008,958 776,860 1,366,891 720,071

20090316 61 10 29 831,309 584,476 1,693,277 953,275

20090315 5 3 92 2,007,496 653,843 57,430 47,557

20090314 1 2 97 228,695 221,965 48,080 40,875

20090313 58 9 33 9,268,570 5,546,437 355,000 263,444

20090312 47 9 44 9,187,750 5,978,887 391,431 274,403

20090311 57 10 34 9,530,332 5,993,967 564,361 507,955

20090310 51 10 39 12,842,679 7,856,675 551,428 475,114

20090309 58 10 32 9,916,761 6,401,931 750,329 562,138

20090308 6 5 90 19,083,218 5,138,443 57,470 54,687

20090307 2 3 95 12,897,700 4,808,327 57,479 47,615

20090306 60 10 30 11,117,580 5,568,263 610,755 533,344

20090305 46 9 46 13,530,229 7,311,263 559,030 351,418

20090304 51 9 41 442,945 365,759 521,445 449,075

20090303 70 12 18 6,502,125 3,858,108 834,985 601,582

20090302 65 13 23 8,945,294 4,989,765 939,699 641,152

20090301 37 15 48 7,670,994 5,422,265 47,061 29,903

this is the page in and page out list, i found that after we restart the system ,the page in increase

and i didn't run the db_verify job ,the page in is reduce. does the db_verify job used the memory?

or to say the db_verify job cause the pagein increase and the disk read slow?

could you help me or give me some suggestion how to deal with it?

thank you very much for your kindly support

Edited by: Xuewen Wang on Mar 31, 2009 3:49 AM