UNIX for Dummies Questions & Answers

Hi,

We have two (2) servers named primary and standby. There is a directory named /db01/archive that we need to keep in-sync.

Files get transferred from primary and standby. Sometimes when we do a failover or when there is a network issue, some files fail to get transferred.

I want to use rsync to sync the files between the primary and the standby. Does rsync checks and ignores files that already exist on each server or does it check if a file has been modified, is therefore newer and then copy it over?

For example, if file01 to file10 are on the primary and file01 to file05 is already on the standby and a network error happens where file06 to file10 were not transferred and I run rsync, can I use rsync to tell it to copy file06 to file10 because they are missing and copy file01 to file05 only if they had been modified and are now newer than the files that are already in the standby?

Similarly, if we failed over from primary to standby and standby has the file11 to file15 and file13 is missing on primary, can I this time run rsync to copy files from the standby to primary where it only copy over file13 and the rest of the other files BUT only if they had been modified and are now newer than the files that are in the primary.

Basically, what I am ultimately hoping to use rsync with is being able to tell it which of the servers is the source and which one is the target whenever I run rsync. I need to be able to do so since because the role between the servers changes when a failover occurs. Is this possible?

Any advice much appreciated. Thanks in advance.

rsync is pretty smart. If all that changes is the meta data (like times), a full copy isn't done if the content is the same. For speed rsync by default looks at time and size to "skip" things it believes don't need to be examined further, but you can disable that with an option.

Thanks, had re-read the man page and I am able to test also using the -in dry run option.

I assume rsync will do a full-on sync between the source and destination. Quite embarrassing, I should have known it will consider the source as the 'master' and the destination as the 'slave'

Anyway, I am using rsync with the -niazv option and it starts as below:

Code:

sending incremental file list
.d..t...... ./
<f+++++++++ testdb_1_1.dbf

Do you know what the .d..t...... ./ mean? Just curious, whether it means this is the root/parent directory and it is skipping it?

If something underneath a directory gets changed in certain ways a directory attribute might be changed. The first char is "." in your example which means nothing was transferred, so there might have been a modification time "t" attribute change made due to some change to something underneath the directory. That "<" on the file means a transfer was made, and so one of the likely culprits of the modification time change to the directory.