Split QSR Backup in chunks

[rohitk1609]

Hi Techies,

I have taken a qlik sense repository backup and size of the backup zip file is more than 15 GB. Now I can transfer so much large file because of its size, Can I split the backed up file in chunks and when I need to restore my Qlik Sense site I can club it back to its original form and use it to restore site.

Thanks in advance,

Rohit

[Gysbert_Wassenaar]

Tools like 7zip for example can split up zip files in multiple parts.

[rohitk1609]

So If I split QSR backup zip file by 7zip and club back with it into 1 , do you think that clubbed file will appropriate to restore a site ?

[Gysbert_Wassenaar]

Yes. Splitting up a zip file would be pretty useless if you can't put the files back together again.

If you like you can first do a md5 checksum on the file that pg_dump spits out and then check that later on the file you retrieve from the zip files. If they're the same then the file is the same and no bits were lost.

[Levi_Turner]

 As @Gysbert_Wassenaar mentioned, most modern compression tools can do splitting (unfortunately the default zipping present in Windows does not). Using parity files (https://en.wikipedia.org/wiki/Parchive) to ensure that you can rebuild the final TAR would make sense.

Honestly, at a higher level, I don't see why a QSR database should be 15GB. What's the result of this query against the QSR:

SELECT schemaname,relname,n_live_tup 
FROM pg_stat_user_tables 
ORDER BY n_live_tup DESC;

?

[rohitk1609]

Thanks levi for adding comment.Big fan of you.
I will execute the your suggested postgrey query by pgadmin and share the result with you but our Qlik site has around 500 apps, 150 around users and 1.5 years old.

second Can you please tell me name the tool or app which I should consider for zipping or unzipping qsr backup ? parchive ?

Third , can you please help me to fix following ticket , I don't have api skill thats why waiting for your input on it

https://community.qlik.com/t5/Qlik-Sense-Deployment-Management/How-to-change-central-node-type-manua...

Thanks in advance
Rohit

[rohitk1609]

Hi Gysbert,
Can you suggest any doc or app which I can use to do md5 checksum, Its a new concept for me.

Thanks,
Rohit

[Levi_Turner]

I have not honestly used any tool for handling TAR files, so I cannot recommend anything specifically for the job. I personally use Bandzip for compression but any modern compression software (WinRAR, 7zip, etc) can handle the splitting. Historically, I like http://www.quickpar.org.uk/ for the parity piece.

[rohitk1609]

Hi Gybert,
Few steps I have move forward:
1.I have taken backup of Qlik site, then reinstall Qlik Sense.
2. I have installed 7-zip and open QSR_Backup.tar file into 7-zip and found 157 dat files.
3. I have created two new folders and transfer halves to one and rest to another.
4. I have created one new folder and transferred both folder .dat files to this one and by
7-zip I archived this new folder to .tar file.
8. i tried to restore this newly .tar file but error is coming as:
"Couldn't find header for file "toc.dat" in tar file"
9. Then I checked the md5 thumbprint of this newly created .tar file(made from ext and split to 2 folder and then clubbed to 1 new folder and converted to tar file) with original QSR_backup.tar file and md5 thumbprint is different.
10. for double check I created one new folder and duplicated it and checked md5 thumbprint is different in this case too.
Can you please help me to know what extactly I am missing ?

Thanks,
Rohit

[Gysbert_Wassenaar]

I don't understand what you're doing. If you use pg_dump to create a backup file you get 1 big tar file. With 7zip you can zip that file and split the archive over multiple parts. All parts are necessary to recreate the original tar file. So why on earth do you put half of those parts in one folder and the other half in another folder? And then you do more weird stuff that's totally unclear to me.

And another thing did you use pg_dump to create a backup of only the QSR database or did you use pg_dumpall which creates a backup that contains all the databases including the QLogs database?