OurDelta

Removal of innodb freeze patch rationale?

Asked by Neil Katin

I was wondering what the rationale for removing the innodb freeze patch in ourdelta-5.0.77-d8?
Was it that the percona hot backup tool was a better solution?

I'm asking because we were using freezing to make consistent binary backups using LVM snapshot and multiple LVM database partitions (our backups were not consistent without freezing first). So we had a use case that required freezing.

We had briefly tried a very early version of xtrabackup; it didn't work for us at the time.

Anyway, this is one sites vote that the innodb freeze feature was useful.

Thanks.

    Neil

Question information

Language:
English Edit question
Status:
Solved
For:
OurDelta Edit question
Assignee:
No assignee Edit question
Solved by:
Neil Katin
Solved:
Last query:
Last reply:
Revision history for this message
Arjen Lentz (arjen-lentz) said : 		#1

IIRC we encountered merging problems so we left it out, indeed considering XtraBackup.
The patch also needs some work in general terms it's incomplete.
Do you want to have a mess with the ourdelta/quilt foo to see if you can get it to merge again?
I'm happy to add it back in for 5.0.

Put up a branch that has it and propose it for merging - I'll take a peek at it and merge to trunk if ok.

Revision history for this message
Neil Katin (d-launchpad-askneil-com) said : 		#2

I'm working on this now, but am slowed down by the fact that the test suite doesn't run after applying all the patches via quilt. Are there known tests that should fail? Or do I have a local implementation issue?

Also, do you have a preferred way of submitting patches? In the absence of any guidance I was going to use the path suggested by the drizzle team, who also use launchpad and bazaar:

    http://drizzle.org/wiki/Contributing_Code

It seemed well thought out and easy for both submitter and reviewer. But if you have a different process that you prefer please let me know.