Existing (corrupted) and new files??

Author Message
Gunnar 11/27/2003 01:45 pm
I have the "Download all" selected as well as "Keep Old Files" (with the date format)

At the moment I`m trying to re-download large rm-files for which the first download resulted in corrupted or missing INDX tables, as well as not all of the file.

Question: Does OE-P still, despite the above selections, try to use the existing old file and negotiate a new download as a continuation of that old file??

I`m using the SR3-1352 at the moment, and just witnessed for a second time how OE-P did not start from the beginning (in seconds) but from , for example, 1500 of 1900 seconds.

Question2: Does OE-P keep information of earlier dowloads (other thann the actual file) which it tries to use for the re-download(s)??
Oleg Chernavin 11/27/2003 03:49 pm
> I`m using the SR3-1352 at the moment, and just witnessed for a second time how OE-P did not start from the beginning (in seconds) but from , for example, 1500 of 1900 seconds.

Yes, it is so, Offline Explorer tries to continue the file download, if it is not complete. I was asked by several users about it - to do that regardless of the Project File Modification check settings.

> Question2: Does OE-P keep information of earlier dowloads (other thann the actual file) which it tries to use for the re-download(s)??

No - it gets that information from the downloaded file only.

Best regards,
Oleg Chernavin
MP Staff
Gunnar 11/28/2003 03:43 am

> Yes, it is so, Offline Explorer tries to continue the file download, if it is not complete. I was asked by several users about it - to do that regardless of the Project File Modification check settings.


OK, thanks, that explains a lot.
To start with I deleted all partial or corrupt files, started a new download, worked pretty good

Then I renamed the corrupted, downloaded again, to compare where things went wrong

And finally I started use the "date-option", and things got just worse.


I believe this is also when I have got strange "double-size-files"??

Gunnar
Gunnar 11/28/2003 03:48 am
Hmm, I wonder if the "download does not start" has to do with attempting to continue from earlier (partial) download??

After deleting all the old corrupt files, downloads, started, went OK, (without rebooting)
Oleg Chernavin 11/28/2003 04:05 am
Yes, maybe so, if it hangs when trying to read data from a corrupted file. Is it possible to reproduce it on some small (less than 1 MB) corrupted .rm file?

Oleg.
Gunnar 12/01/2003 07:45 pm
Not easy, the files I attempt to get, and have problems with, are 10-200MB.

However, I just experienced some "non-starts" (but non-start URLS were added to the end of the queu??), suspended to file, exited and got "blue screen" exception at 004024FC reading 59B2F014, clearly something had gone badly wrong. (has happened a couple of times, althooughh I did noot write down the exception address)

Resuming (from file) seems to work??

Note, I`m only doing a handful of rm-files, but size 100-200MB, pasting URLs into list, project properties, but it obviously takes hours.

Ref INDX tables, I could hex-edit-cut-copy-save some examples of weird, missing, wrong position INDX tables, as soon as I get some systematic results. (smaller than the original 100-200MB files)

Gunnar


> Yes, maybe so, if it hangs when trying to read data from a corrupted file. Is it possible to reproduce it on some small (less than 1 MB) corrupted .rm file?
>
> Oleg.
Oleg Chernavin 12/03/2003 10:10 am
> However, I just experienced some "non-starts" (but non-start URLS were added to the end of the queu??), suspended to file, exited and got "blue screen" exception at 004024FC reading 59B2F014, clearly something had gone badly wrong. (has happened a couple of times, althooughh I did noot write down the exception address)

Is this reproduceable? If so, how I can repeat it?

> Note, I`m only doing a handful of rm-files, but size 100-200MB, pasting URLs into list, project properties, but it obviously takes hours.

How do you do that?

> Ref INDX tables, I could hex-edit-cut-copy-save some examples of weird, missing, wrong position INDX tables, as soon as I get some systematic results. (smaller than the original 100-200MB files)

OK. Thank you!

Oleg.
Gunnar 12/03/2003 02:56 pm
> > However, I just experienced some "non-starts" (but non-start URLS were added to the end of the queu??), suspended to file, exited and got "blue screen" exception at 004024FC reading 59B2F014, clearly something had gone badly wrong. (has happened a couple of times, althooughh I did noot write down the exception address)
>
> Is this reproduceable? If so, how I can repeat it?

Got another "non-start" situation, this time the exception was at 0040247C.

That is, after noticing that the next rm-file in the que does not start, I suspend to a file, but have got these two error-messages (plus earlier something general about a pointer problem).

Restarting OE-P and continuing from the suspended file works OK.


>
> > Note, I`m only doing a handful of rm-files, but size 100-200MB, pasting URLs into list, project properties, but it obviously takes hours.
>
> How do you do that?

I have my own program which collects the rm,ram or smi URLS as well as checks if I already have them on my hard disk. The ones I decide on downloading are put into a txt-file, from which I copy-paste into the "properties" of a (the same, again and again) OE-P project.

After that I "start" the OE-project.

Usually between 2 to 6 URLs.

That is, I do not use any of OEs capabilites to parse links and find new links. (except for ram and smi files, for which I`m still trying to figure out what OE-P does, have been using URLSnoop to find them "manually")

The problem is that all these files are 50-300 MB, 0.5 to 3 hours of 100kbps rm-streams, takes some time to repeat some particular project-download.

However, from the 3-4 "non-start" cases I have observed, it seems to happen after "some time".

That is, these "exit-and-exception" non-starters can be handled by suspending to a file and restarting OE, then resuming from file.

There seems to be another "non-start" problem for just particular links in the que?? That is, after timeout the next (rtsp-) rm-link starts normally?? (however, I have not seen them lately, maybe because I only do one RTSP channel, 90s delay inbetween and delete "corrupted" attempts??)

> > Ref INDX tables, I could hex-edit-cut-copy-save some examples of weird, missing, wrong position INDX tables, as soon as I get some systematic results. (smaller than the original 100-200MB files)
>
> OK. Thank you!


Need to find some "typical ones", seems like these ones solved my "original" problems:

- download only one file at a time (maybe that temp-directory problem??)
- insert 90s delay between downlaods (time to move the 200MB tmp-file and finalize it??)
- delete "problem files" before downloading them again??

Note, all of this only applies to having an OE-project with just 3-5 (maybe 10) rtsp-rm links pasted into the project "URLs" list (only audio-video files crossed, level=1, download all files, etc)

That is, OE-P seem to not be targeting the way I use it...

Ouch, another "non-start" case... this time after a OE-restart and only pasting two links to be downloaded.. need to wait for some additional of those 90s delays I have specified....

...
...


OK, timeout, restarts my 90s delay, attempting the same link
changing the delay to just 9 seconds, started realether monitoring on another networked PC
..

Ouch, copy-pasting the rtsp-link into the realvideo player does not get the stream going...
Trying the same link on another (this) PC, original RealVideo Player.. also a "non-starter"..

OK, need to reboot...

Hey, nobody promised it would be easy...

Gunnar

The non-start link, if of any interest

rtsp://video.c-span.org/60days/brit120303.rm









Gunnar 12/03/2003 02:56 pm
> > However, I just experienced some "non-starts" (but non-start URLS were added to the end of the queu??), suspended to file, exited and got "blue screen" exception at 004024FC reading 59B2F014, clearly something had gone badly wrong. (has happened a couple of times, althooughh I did noot write down the exception address)
>
> Is this reproduceable? If so, how I can repeat it?

Got another "non-start" situation, this time the exception was at 0040247C.

That is, after noticing that the next rm-file in the que does not start, I suspend to a file, but have got these two error-messages (plus earlier something general about a pointer problem).

Restarting OE-P and continuing from the suspended file works OK.


>
> > Note, I`m only doing a handful of rm-files, but size 100-200MB, pasting URLs into list, project properties, but it obviously takes hours.
>
> How do you do that?

I have my own program which collects the rm,ram or smi URLS as well as checks if I already have them on my hard disk. The ones I decide on downloading are put into a txt-file, from which I copy-paste into the "properties" of a (the same, again and again) OE-P project.

After that I "start" the OE-project.

Usually between 2 to 6 URLs.

That is, I do not use any of OEs capabilites to parse links and find new links. (except for ram and smi files, for which I`m still trying to figure out what OE-P does, have been using URLSnoop to find them "manually")

The problem is that all these files are 50-300 MB, 0.5 to 3 hours of 100kbps rm-streams, takes some time to repeat some particular project-download.

However, from the 3-4 "non-start" cases I have observed, it seems to happen after "some time".

That is, these "exit-and-exception" non-starters can be handled by suspending to a file and restarting OE, then resuming from file.

There seems to be another "non-start" problem for just particular links in the que?? That is, after timeout the next (rtsp-) rm-link starts normally?? (however, I have not seen them lately, maybe because I only do one RTSP channel, 90s delay inbetween and delete "corrupted" attempts??)

> > Ref INDX tables, I could hex-edit-cut-copy-save some examples of weird, missing, wrong position INDX tables, as soon as I get some systematic results. (smaller than the original 100-200MB files)
>
> OK. Thank you!


Need to find some "typical ones", seems like these ones solved my "original" problems:

- download only one file at a time (maybe that temp-directory problem??)
- insert 90s delay between downlaods (time to move the 200MB tmp-file and finalize it??)
- delete "problem files" before downloading them again??

Note, all of this only applies to having an OE-project with just 3-5 (maybe 10) rtsp-rm links pasted into the project "URLs" list (only audio-video files crossed, level=1, download all files, etc)

That is, OE-P seem to not be targeting the way I use it...

Ouch, another "non-start" case... this time after a OE-restart and only pasting two links to be downloaded.. need to wait for some additional of those 90s delays I have specified....

...
...


OK, timeout, restarts my 90s delay, attempting the same link
changing the delay to just 9 seconds, started realether monitoring on another networked PC
..

Ouch, copy-pasting the rtsp-link into the realvideo player does not get the stream going...
Trying the same link on another (this) PC, original RealVideo Player.. also a "non-starter"..

OK, need to reboot...

Hey, nobody promised it would be easy...

Gunnar

The non-start link, if of any interest

rtsp://video.c-span.org/60days/brit120303.rm









Gunnar 12/03/2003 03:12 pm
Wrong alarm, that link was a "non-starter" because of other reasons, the server or something else.

That is, rebooted, pasted into a regular Relavideo player, opened the actual htm-page, Realvideo says "nothing exists", but other rm-links on the same page works ok.

Exited the OE (on the other PC) and it did not crash (in comparison to other "non-start" problems which I have solved by "suspend to file", exiting and resuming.

From what I quickly scanned through of the etherreal monitoring of the traffic, it looked like the target-server just did not respond when the actual streaming should have started. OE retried again and again after the time-out and delay I have specified..

Things happen...

Gunnar