[Freeipa-users] Split backup actions in stop - backup - start commands

[Matt .]

Hi guys,

As I'm using burp for backup I get the feeling it fails obt eh
ipa-backup proces itself when runned as a pre_script. I think it waits
for some exitcode or already gets it before the real backup of IPA has
been finished.

I'm checking this out as burp also outputs messages as errors because
it just does it that way.



2016-02-18 16:08 GMT+01:00 Rob Crittenden <rcritten at redhat.com>:
> David Kupka wrote:
>> On 17/02/16 10:47, Matt . wrote:
>>> Hi David,
>>>
>>> I have tested your way out and it seems to be OK.
>>>
>>> The reason why I need this was is so I can perform a stop and
>>> ipa-backup before I start my backup to my backupserver. (pre-command).
>>>
>>> If I use ipa-backup directly it errors between the stop of ipa and the
>>> actual ipa backup. I need to check that out further.
>>>
>>> An ipactl start is not needed it seems as the ipa-backup command seems
>>> to start ipa at any time again.
>>>
>>> Do you understand/agree here ?
>>
>> Hello Matt,
>>
>> unfortunately I don't understand. The backup procedure AFAIK should work
>> like this:
>>
>> # ipa-backup && rsync -r /var/lib/ipa/backup/ backup.example.test:/ipa/
>>
>> You ca run it manually or place it into the crontab or use it in your
>> orchestration system.
>> It will backup the ipa server with necessary stop and start and then
>> copy the new backup to the backup server.
>>
>> Still I don't see the need for stopping the server manually.
>>
>> ipa-backup calls "ipactl start" [0]. If you remove the else branch it
>> will not start the server.
>>
>> [0
>> ]https://git.fedorahosted.org/cgit/freeipa.git/tree/ipaserver/install/ipa_backup.py#n316
>
> I've also been wondering about the request to remove the stop/start. The
> stop should be idempotent. Is an error being thrown?
>
> rob
>