Archive question, please help ~

[raygrrr]

Dear experts,

I am wondering is it possible to archive the files in my openkm server, say, over one year old files? I was doing a lot searching, but they all talking about how to do the whole backup, or incremental backup also not quite what I was looking for.

So the only solution I could come up is that to run two instances of openkm, and use webservice to move files(over one year old) from the current openkm instance to the archive openkm instance, is that possible or there are better way to do this?

Why I want to do this is because my openkm is getting bigger and bigger, I wanna keep it under certain size while not to physically remove old files.

Many thanks!
Ray

[jllort]

Archiving can be done in several ways, one solution is package all document ( with vesion, metadata etc... ) into a zip ( or other format ) and store it with some metadata what can be used in future to recover information.

Here must take in mind several things before starting:
1- what happens with security in openkm when use for archived ( normally is not applied the same security )
2- who can access it
3- and if it's needed what will be used for doing search.

The idea is package the information as is in openkm ( consider all openkm as an object what you want to store in some fomat -> zip ) and then what information is needed to recover it on future. A possible scenario is do export process to server ( document with history ) then pack all into a zip, and then import into other openkm with metadata.

If you take a look at our import / export format, and elegant way for doing it is when you got zip file, create the .okm file ( really it's a simply json file ) and then use the other openkm to import ( in this scneario is not neeed webservices, has not lost any information, and in future if it's needed zip file can be restored into openkm ).

In openkm professional version we can store documents outside openkm ( with this format ) and as a part of document export is possible create record what only contains metadata.

These are only some ideas about how doing it. I do not know if your problem is repository size or number of objects stored in openkm. Is not exactly the same problem.

[raygrrr]

Hi jllort,

Many many thanks for your detailed answer!!!

Regarding your questions:
1- what happens with security in openkm when use for archived ( normally is not applied the same security )
- we are not really concern about this security issue as it's in our own server (internal network).
2- who can access it
- same as above
3- and if it's needed what will be used for doing search.
- we will not directly search from archive file.


Our current plan is:
1) just use one openkm instance and archived the whole package (so including the mysql db) first
2) then delete the old files (say over one year old) to keep the current openkm size small while still can have the full backup.

Do you think above solution is OK? thanks!

Thanks again,
Ray

[jllort]

First think on restrictions what must accomplish your historical, who can access, what you'll looking for, etc. when you got this scenario well defined then you can thing on how getting it.- one option could be replicate security, etc. before arrive this detail is necessary define your scenario, your needs, and then when you know it, then can decide what's better for your, is not easy suggest something, because almost problems have not only one way to solve it.

A good option is to export, before deleting. About incremental backups you can do it from exported data, for example some folder in server where you daily export purged documents - historical logic -.

[ambrose]

I have registered a property group, and added metadata to a number of files successfully as a test. However, I cannot then search by properties as the "Add Property Group" in the metadata search tab is grey/does nothing upon clicking.

[pavila]

Which OpenKM version are you using?