SAP OS Upgrade from AIX to Linux – Part2
In the first blog of the series we have seen the export preparations and table splitting process. In this blog we will see how to export AIX DB.
1. Start again SWPM process –> Select Database Instance Export like shown below and click Next.
2. Specify Profile Directory.
3. Specify Database SID and host name.
4. We have downloaded the kernel from SAP market place and will use this kernel for this process. So select checkbox Use Dedicated Kernel for System Copy in the below screen.
5. Specify the kernel path.
6. Specify the location for Database Export.
7. Specify Target Database and select Split STR Files.
8. Now let’s go ahead and create a shared folder called SQL.
9. Login into ABAP stack and run program SMIGR_CREATE_DDL from SE38 transaction. In the selection screen specify the folder we have created above.
10. Executing the above program will give us the below result.
11. Now go back to export process and specify the SQL folder which we created above.
12. Select New Export from Scratch.
13. Specify Use database-specific implementation option.
14. Specify the parameters for splitting process.
15. Choose Little Endian and specify 72 for number of parallel jobs.
How we got 72 is 3 times the number of lcpu. In my current system we can found out at OS level using lparstat. Not necessarily we have to use this way, the number we gave doesn’t over the kill the system and at the same time if we give less number it will take more time to export.
16. Specify Key Phrase.
17. Specify the Package Split size. In my case we have given as 300M.
18. Specify the number of Parallel jobs.
19. Specify No for Database Stats.
20. Check the parameters specified so far and click on Next.
21. Once the export starts it will continue and will stop at this process if the current system is running.
22. Now shutdown all the applications except DB and say OK in the above screen.
23. If we don’t run into any issues system will export the DB successfully.
This completes the second blog in this series. What we have learnt in this part is how to export the DB in the existing AIX landscape. In the next part3 we will see on how to install ascs, scs, servers.
Источник
Tools for Oracle DB migration from AIX to Linux
My colleague running Oracle Database (11g) in AIX and they would like to move this DB to RHEL. I already found Link. However I would like to check if someone have already migrated or used any other best tools.
2 Answers 2
you have several options. As pointed out before, Oracle Data Pump is the easiest approach. It would lift you from every version >=10g upwards (or even back when you use the VERSION= parameter).
The caveat is: Size of the database — and your downtime requirements.
In terms of larger databases, Transportable Tablespaces is the usual choice. More work as you will have to rebuild meta information such as synonyms, view, plsql, sequences etc — and in your case you’ll have to either CONVERT the tablespaces as you are coming from a Big Endiann platform and going to a Little Endiann. DBMS_FILE_TRANSFER could assist you here as it can restore and covert at the same time whereas RMAN will need a 2-phase operation with staging space for it.
You can speed up transportable tablespaces with RMAN Incremental Backups to avoid most of the copy/convert time. And you can ease it with Full Transportable Export/Import (minimum source: 11.2.0.3 — minimum destination: 12.1.0.1) where Data Pump does the manual work of transportable tablespaces.
And of course there are other techniques such as Create-Table-As-Select or Insert-Append-Select options via Database Links and such.
Just see the big slide deck «Upgrade / Migrate / Consolidate to 12.2» for customer examples — and the «Migrate >230Tb in
Источник
Migrating DB2 UDFs from AIX to Linux (RHEL)
I successfully compiled some UDFs on 64-bit Linux that were originally compiled on AIX
Linux is running IBM DB2 v9.5
Here’s the function that sometimes returns for some values
And here’s the data:
Does anybody have any ideas?
My guess is that this if statement:
is returning true on some values. Could it be something related to 64bit?
Also the couts do not print in the DB2 output
1 Answer 1
What are the rules for converting a portion of the binary zDate to a usable date? You could probably come up with an equivalent conversion function in SQL and avoid the external compiled UDF entirely. The HEX function will convert the binary value to a hex string, which you can then iterate through with SUBSTR to build your integer value.
As for the cout messages, I don’t think that cout is going to be of much use inside an external UDF. You’ll have better luck saving your debugging info into a local file.
Related
Hot Network Questions
Subscribe to RSS
To subscribe to this RSS feed, copy and paste this URL into your RSS reader.
site design / logo © 2021 Stack Exchange Inc; user contributions licensed under cc by-sa. rev 2021.10.8.40416
By clicking “Accept all cookies”, you agree Stack Exchange can store cookies on your device and disclose information in accordance with our Cookie Policy.
Источник
Linux
AIX 6.1 to Linux 7.2 migration
Hi,
recently we have migrated our current AIX server to Linux, we have lot of shell script, few of them are FTP scripts.
we have copied the complete AIX file system to linux 7.2 as it is.
could you please highlight what are the things we need to look into it .
in AIX we are using .netrc to store all FTP destination details servername,userid, password . etc, will .netrc work in linux environment , if not what is the replacement of this file.
also please highlight the areas which we really need to look in terms after migration , currently we have just migrated our TEST box.
Thank you in advance.
By ‘. copied the complete AIX file system to linux. ‘ can I assume that you mean the application filesystem? Overwriting operating system parts could cause unpredictable issues.
You might just find that the ftp client and server are not installed by default on your linux system (you don’t say which supplier, which makes it tricky)
You might find vsftpd gives you this service, but you have to install it. You may also need to install an ftp client. What OS are you moving to? Common Linux providers are RedHat, CentOS, Debian, Mint, etc. and they all differ a bit on how you do things. Pasting the output from uname -a (in CODE tags) would help us here.
As far as I know, the .netrc files should still be valid with any ftp client.
Remember that the client opens the connection to the server so you need to know which side is in charge. You may only need one of these packages.
Источник
UNIX for Dummies Questions & Answers
AIX to Linux command difference
Use ksh instead bash . You are probably used to ksh anyway when you come from AIX and it is available on every Linux since some time now. When using ksh you have the shell-internal » print » for output. It allows to explicitly end options by using a single dash:
will output «foo bar» to stderr (-u2) but:
will write «-u2 foo bar» to stdout , because «-u2» is no longer considered an option.
Per default i use it that way, especially if i do not know what a variable i want to display contains:
Even if «$var» contains options to «print» they will be ignored.
I hope this helps.
Ok. so in general we have an old SysV vs. Berkley. or perhaps a ksh vs. Berkley-ism.
I too was somewhat displeased with the Linux choice of a BSD-ish like echo since it tried to be more POSIX which made is smell more SysV like (Linux is sort of its own blend).
So the niceness of echo «first line\nsecond line» that you had in HPUX, Solaris, etc.. well, you had to do echo -e «first line\nsecond line» in Linux. Now. with more contemporary versions of ksh, they’ve added a «do nothing» option of -e, which means for any system running ksh93 (maybe even something less) but not not a really old ksh, you could use -e to make scripts seemingly portable.
But. maybe you plan to run ksh93 on your Linux host.
Then, you can run the following:
Sadly, even getconf syntax has changed over the years, for the command getconf UNIVERSE = att, you may have to use a minus instead of an equals.
Now. with regards to that latter solution, if you whole system runs in ksh and uses a non-systemd SysV style init, you could put that stuff there, but of course I can pretty much guarantee you that your distro producers didn’t have portability in mind when they wrote their init scripts. So. it will likely render your system useless to try change away from bash.
There’s like a KSHENV var that might be able to pull such things in for any ksh script. it’s been too long.. anyhow, might get you pointed in the right direction.
Источник
