You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@subversion.apache.org by void pointer <rc...@gmail.com> on 2009/04/01 16:55:14 UTC
Re: "Decompression of svndiff data failed"
Weird, now I'm getting this:
Traceback (most recent call last):
File "./fsfsverify.py", line 1134, in <module>
options.dumpWindows)
File "./fsfsverify.py", line 571, in verify
svndiff = Svndiff(f, self.length)
File "./fsfsverify.py", line 461, in __init__
(self.startingOffset)
__main__.InvalidSvndiffHeader: Invalid svndiff header at offset 15639751
I got this when I was running the fsfsverify.py script without the -f
switch.
On Tue, Mar 31, 2009 at 2:49 PM, Watson, Chris <cw...@ptc.com> wrote:
> I just use a shell script:
>
>
>
> REPO=$1
>
> REV=$2
>
>
>
> /opt/ptc/bin/fsfsverify.py -f /opt/ptc/subversion/sets/$REPO/db/revs/$REV
>
> /opt/ptc/bin/fsfsverify.py /opt/ptc/subversion/sets/$REPO/db/revs/$REV
>
> until [ $? -eq 0 ]
>
> do
>
> /opt/ptc/bin/fsfsverify.py -f /opt/ptc/subversion/sets/$REPO/db/revs/$REV
>
> /opt/ptc/bin/fsfsverify.py /opt/ptc/subversion/sets/$REPO/db/revs/$REV
>
> done
>
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=1065&dsMessageId=1508866
To unsubscribe from this discussion, e-mail: [users-unsubscribe@subversion.tigris.org].
Re: "Decompression of svndiff data failed"
Posted by John Szakmeister <jo...@szakmeister.net>.
On Wed, Apr 1, 2009 at 1:44 PM, John Szakmeister <jo...@szakmeister.net> wrote:
> On Wed, Apr 1, 2009 at 12:59 PM, void pointer <rc...@gmail.com> wrote:
>> Also, when I run svnadmin verify, I get this:
>> svnadmin: Corrupt representation '54 15639745 1098147 2113536
>> 7ce69b35ce9d74eac37901010cfb7c1f'
>> svnadmin: Malformed svndiff data in representation
>
> Right. Some people are unfortunate enough to have multiple errors in
> your rev. I've had to run it upwards of 14 times to fix all the
> errors. And, there are some that fsfsverify won't fix (it'll attempt,
> but ultimately fails).
>
> To answer several of your questions:
> 1) Why doesn't fsfsverify fix all the errors? Because it really
> can't proceed once it finds the first corruption. It could attempt to
> fix and reparse, but I prefer to be in control of how many times the
> tool is run.
>
> 2) Are you screwed? Probably not. It may take several attempts
> before fixing all the errors. Additionally, there may be errors in
> other revisions. Is svnadmin failing in the same revision as before?
>
> I can probably take a quick look at the revision, but you need to get
> it to me ASAP as I'm at a conference, and will be leaving in a short
> while. If the file is big, post it somewhere and give me a url,
> otherwise email it directly to me.
I should add: send me the original file (the one you backed up before
running fsfsverify).
-John
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=1065&dsMessageId=1509325
To unsubscribe from this discussion, e-mail: [users-unsubscribe@subversion.tigris.org].
Re: "Decompression of svndiff data failed"
Posted by John Szakmeister <jo...@szakmeister.net>.
On Wed, Apr 1, 2009 at 12:59 PM, void pointer <rc...@gmail.com> wrote:
> Also, when I run svnadmin verify, I get this:
> svnadmin: Corrupt representation '54 15639745 1098147 2113536
> 7ce69b35ce9d74eac37901010cfb7c1f'
> svnadmin: Malformed svndiff data in representation
Right. Some people are unfortunate enough to have multiple errors in
your rev. I've had to run it upwards of 14 times to fix all the
errors. And, there are some that fsfsverify won't fix (it'll attempt,
but ultimately fails).
To answer several of your questions:
1) Why doesn't fsfsverify fix all the errors? Because it really
can't proceed once it finds the first corruption. It could attempt to
fix and reparse, but I prefer to be in control of how many times the
tool is run.
2) Are you screwed? Probably not. It may take several attempts
before fixing all the errors. Additionally, there may be errors in
other revisions. Is svnadmin failing in the same revision as before?
I can probably take a quick look at the revision, but you need to get
it to me ASAP as I'm at a conference, and will be leaving in a short
while. If the file is big, post it somewhere and give me a url,
otherwise email it directly to me.
-John
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=1065&dsMessageId=1509314
To unsubscribe from this discussion, e-mail: [users-unsubscribe@subversion.tigris.org].
Re: "Decompression of svndiff data failed"
Posted by void pointer <rc...@gmail.com>.
Also, when I run svnadmin verify, I get this:
*svnadmin: Corrupt representation '54 15639745 1098147 2113536
7ce69b35ce9d74eac37901010cfb7c1f'*
*svnadmin: Malformed svndiff data in representation*
On Wed, Apr 1, 2009 at 11:55 AM, Robert Dailey <rc...@gmail.com> wrote:
> Weird, now I'm getting this:
> Traceback (most recent call last):
> File "./fsfsverify.py", line 1134, in <module>
> options.dumpWindows)
> File "./fsfsverify.py", line 571, in verify
> svndiff = Svndiff(f, self.length)
> File "./fsfsverify.py", line 461, in __init__
> (self.startingOffset)
> __main__.InvalidSvndiffHeader: Invalid svndiff header at offset 15639751
>
> I got this when I was running the fsfsverify.py script without the -f
> switch.
>
> On Tue, Mar 31, 2009 at 2:49 PM, Watson, Chris <cw...@ptc.com> wrote:
>
>> I just use a shell script:
>>
>>
>>
>> REPO=$1
>>
>> REV=$2
>>
>>
>>
>> /opt/ptc/bin/fsfsverify.py -f /opt/ptc/subversion/sets/$REPO/db/revs/$REV
>>
>> /opt/ptc/bin/fsfsverify.py /opt/ptc/subversion/sets/$REPO/db/revs/$REV
>>
>> until [ $? -eq 0 ]
>>
>> do
>>
>> /opt/ptc/bin/fsfsverify.py -f
>> /opt/ptc/subversion/sets/$REPO/db/revs/$REV
>>
>> /opt/ptc/bin/fsfsverify.py /opt/ptc/subversion/sets/$REPO/db/revs/$REV
>>
>> done
>>
>
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=1065&dsMessageId=1508895
To unsubscribe from this discussion, e-mail: [users-unsubscribe@subversion.tigris.org].