On 2010-09-24, at 5:23 PM, Daniel Shahaf wrote:
> Luis Rojas wrote on Fri, Sep 24, 2010 at 15:16:17 -0400:
>>
>> On 2010-09-24, at 1:27 PM, Daniel Shahaf wrote:
>>
>>> Luis Rojas wrote on Fri, Sep 24, 2010 at 11:42:07 -0400:
>>>> suddenly the revision in the copy started passing verification, so we
>>>> replaced that file in the original repository and all was well...
>>>
>>> Did you compare the contents of the file in the original repository with
>>> the contents of the file you replaced it with?
>>
>> what do you mean? the file i replaced was the actual revision file on
>> /repository/db/revs/
>
> I understand this. You said that you did:
>
> mv RepoCopy/db/revs/N Repo/db/revs/N
>
> I'm asking if you compared these two files before you replaced one with
> the other.
>
Nope, i cant say i did, since the revision is a binary file i didn't saw the point in comparing ( not that i have anything to diff a couple of binary files ).
>> but the file that was on that commit was the
>> same. ( after we restored it we compared it against the email the hook
>> script sends )
>>
>>>
>>>> for 2 hours or so, after the same developer committed
>>>> again a test file, we started getting a new error "Revision file lacks
>>>> trailing newline".
>>>>
>>>
>>> Can you rule out hardware (disk) problems?
>>>
>>> What processes were accessing the repository around the time of those
>>> unfortunat ecommits? (other commits? revprop changes? svnadmin?)
>>>
>>
>> We actually suspect the disk, but we don't have enough evidence to
>> suggest the disk is faulty, we are a small development team, and is
>> very rarely we submit at the same time, i am not saying it cant happen
>> but is very unlikely. also, as far as i know, no other process where
>> accessing the repository at the time.
>>
>
> "No other process" --- did you mean: no other process except for the
> post-commit hook script?
>
Exactly, no other process other than the post-commit hook, although our post-commit only reads info from the repository ( svnlook ) to integrate with mantis, basically read commit info, send it to mantis... that's it.. nothing fancy... on top of that, is been working fine for years.
>>
>>>>
>>>
>>> Thanks,
>>>
>>> Daniel
>>
>> Luis R. Rojas
>> SysAdmin
>> Rogue Research Inc
>> T: (514) 284-3888
>> F: (514) 284-6750
>> luis_at_rogue-research.com
>>
>>
>>
Luis R. Rojas
SysAdmin
Rogue Research Inc
T: (514) 284-3888
F: (514) 284-6750
luis_at_rogue-research.com
Received on 2010-09-27 16:06:49 CEST