Incremental Online Backup much smaller after upgrading to 11

Posted by bernhardkraml on 31-Oct-2018 02:40

I'm a little bit unsure about the quality of my backup on the new system.

I think I read time ago about a Problem with some Counter in the incremental mechanism causing a backup with not all changed blocks in the database?

I'm not sure, but on the old system (10.2b) my online incremental backup with -io 5 and -com was always around 3 GByte. Now the Incremental Backup is only around 160 MByte

Tha database has an size of ~140 GByte

Migration was done with backup (10.2b) on the old system

on the new system prostruct and prorest with 10.2b utilities (I installed the 10.2b version)

then with 11.7.2 utilities:

Proutil conv1011.

Who has ideas about the cause for this size-reduction?

I just did a restore with the full-backup from last weekend and the incremental from yesterday : there was no error message.

By the way, also the full backup has an restored size 78 Gig on 10.2b 70 Gig on the new one??

-bernhard

Posted by Paul Koufalis on 31-Oct-2018 07:47

This is expected behaviour. In or around 11.3, new functionality was added to only backup active BI clusters during an online backup. Prior to that, the entire BI was always backed up during probkup online.

You can reproduce the old behaviour by adding -bibackupall to your probkup online command.

As to your 2nd question, I would be curious to hear which files are smaller. Is it the BI file(s)?

All Replies

Posted by cjbrandt on 31-Oct-2018 07:28

Does a dbanalys show different numbers from 10.2B to 11.7.2 ?

Posted by Paul Koufalis on 31-Oct-2018 07:47

This is expected behaviour. In or around 11.3, new functionality was added to only backup active BI clusters during an online backup. Prior to that, the entire BI was always backed up during probkup online.

You can reproduce the old behaviour by adding -bibackupall to your probkup online command.

As to your 2nd question, I would be curious to hear which files are smaller. Is it the BI file(s)?

Posted by mollyfed on 31-Oct-2018 07:48

Hmmm... I wonder if this (knowledgebase.progress.com/.../P138656) is what you were refering to and if you are actually getting a fix for the issue in the later release that you are using?

The original Kbase doesn't mention that a fix was found but perhaps it has been and no one has updated the kbase.

To be honest, if you can confirm that your backup and restore are fine, you probably are fine!

JTMC

Molly

Posted by bernhardkraml on 31-Oct-2018 09:25

@mollfed

I cant open the link knowledgebase.progress.com/.../P138656)

by the way, is there any method to retrieve KB-Articels when knowing P138656?

The Restore (full and last incremental, i have -io 5) gave no error.

I checked some information I knew I entered yesterday before the backup run  - I also found this information.

@ Paul Koufalis

This new behavior would make sense.

My BI-File is ~ 2Gig (fixed File with second variable extent) , there is almost no activity at night - so there shouldn't be very much active BI cluster.

And this is a Payroll-Database - there wasn't much change on the last 2 days.

So I think is all OK.

As long there is no problem with not having backup-ed the complete BI clusters.

Wasn't there some problem with this?

Posted by Rob Fitzpatrick on 31-Oct-2018 09:27

> Wasn't there some problem with this?

It was broken in 11.3.x and 11.4, fixed in 11.5+.

This thread is closed