Chris Withers wrote at 2007-3-22 08:43 +:
>Dennis Allison wrote:
>And that I'm lazy and really want to be able to do:
>
>python rollback.py 2007-03-21 09:00
You have been told that you can specify a "stop" time
and the storage will "stop" at the given time.
Thus, you look at the code how this
Dennis Allison wrote:
The ZODB is an append only file system so truncating works just fine.
Yup, but it's finding the location to truncate back to that's the
interesting bit.
And that I'm lazy and really want to be able to do:
python rollback.py 2007-03-21 09:00
You can use any
of the sta
Jens Vagelpohl wrote:
On 21 Mar 2007, at 17:29, Chris Withers wrote:
I'm hoping for some means to just lop transactions off the end of the
Data.fs until I get to the point in time I want...
Why don't you just record the exact size of your Data.fs before starting
your migration and then trunca
Jim Fulton wrote at 2007-3-21 10:06 -0400:
> ...
>
>On Mar 21, 2007, at 9:59 AM, Tres Seaver wrote:
>
>> -BEGIN PGP SIGNED MESSAGE-
>> Hash: SHA1
>>
>> Jim Fulton wrote:
>>> On Mar 21, 2007, at 6:41 AM, Chris Withers wrote:
>>>
Hi All,
Is there any existing method or script f
Chris,
The ZODB is an append only file system so truncating works just find. I
have used truncation occasionally to fix a coddled ZODB. You can use any
of the standard tools (fsrecover, for example) to find the transactions
and their dates. The position in the file is part of the fsrecover
out
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 21 Mar 2007, at 17:29, Chris Withers wrote:
I'm hoping for some means to just lop transactions off the end of
the Data.fs until I get to the point in time I want...
Why don't you just record the exact size of your Data.fs before
starting you
Laurence Rowe wrote:
- Should this create a new FileStorage? Or should it modify the
existing FileStorage in place?
Probably create a new one (analogous to a pack). Seems safer than
truncating to me.
Nah, this is working on a copy of production data, not the real thing.
Disk space is an iss
Jim Fulton wrote:
- Should this create a new FileStorage?
No.
Or should it modify the existing
FileStorage in place?
Yes.
- Should this work while the FileStorage is being used?
No.
- Should this behave transactional?
Not sure what you mean... in terms of being able to "undo the u
Tres Seaver wrote:
1. Open the existing file in time-travel mode (readonly, so that it
loads no transactions later than the given point). The 'stop'
parameter to the FileStorage initializer is the one which
triggers time-travel.
2. Open a new, writable filestorage.
3. Run
Jim Fulton wrote:
I wasn't asking about implementation.
Here are some questions:
- Should this create a new FileStorage? Or should it modify the existing
FileStorage in place?
Probably create a new one (analogous to a pack). Seems safer than
truncating to me.
- Should this work while t
On Mar 21, 2007, at 9:59 AM, Tres Seaver wrote:
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Jim Fulton wrote:
On Mar 21, 2007, at 6:41 AM, Chris Withers wrote:
Hi All,
Is there any existing method or script for rolling back a ZODB
(filestorage-backed in this case,
Back end to what?
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Jim Fulton wrote:
> On Mar 21, 2007, at 6:41 AM, Chris Withers wrote:
>
>> Hi All,
>>
>> Is there any existing method or script for rolling back a ZODB
>> (filestorage-backed in this case,
>
> Back end to what?
>
>
>> if that makes it easier) to
12 matches
Mail list logo
