Problems with uploads of bigger files

[LHammonds]

buddah and Dark0ne have been working on this all day. I've seen their chats in the Staff Area.

 

LHammonds

[Yossarian22]

I've seen, buddah has uploaded my file today, so hopefully they had some success. Thanks for uploading anyway :thanks:

[buddah]

We have checked the site, several times I uploaded the same file to your page several times and then deleted them. It is not an issue specific to the site that we have been able to determine.

 

What I suspect is a corruption within the ATM network as it will bomb out occasionally for no reason. We ran more tests today and all uploads were successful with no dropouts/corrupted files. I monitored a steady upload rate with no errors/corrupted blocks hcec header errors. Downloaded the files and ran a CRC check against the originals and tested with 7 and winrar and the archives were good.

 

It has been a few days since I had an upload bomb out...so the problem may now be corrected. So please give it a try and let us know what happens.

 

A supposition on my part is there was a bad hop that is showing up on a few connections....they are marginal trunks within the internet itself they will be good for most users of the net. But when trying to pass large amounts of data for a single file they actually fail completely due to packet corruption and loose the header data of the octet packet and break the connection to the server.

 

I shall wait to hear if your uploads work now.

 

Buddah

[Arthmoor]

Whatever you guys have been doing, it seems to have greatly improved my chances of being able to upload. Rather than failing consistently at 2MB each time, it's getting up to about 10-12MB each time now. Better than before, but not good enough for the larger stuff.

 

I don't suppose there's any chance at all of getting upload resume at some point? It would sure help for situations like this.

[Arthmoor]

Interesting bit of info if you're still looking for answers.

 

I just sent up a 32MB file. Only it wasn't quite in the normal manner. I used the normal upload form to start the process, then as usual it failed. I decided to see what would happen if I told Firefox to retry. It prompted to resend the POST data. It sat there in that state for quite some time while the file uploaded, but it did eventually make it through.

 

During all this I was monitoring the TCP/IP connections using a task manager extension called "Prio". During the failed upload, there was traffic to the upload server, but also two extra connections opened up that were thrashing data back and forth for awhile. They stopped doing so, and very shortly after that is when the main upload failed. I think those two extra data connections were for the upload status indicator on the page.

 

The retry/resubmit action only reopened a single TCP/IP connection back to the upload server which had no real trouble other than being a bit slow, but since you're in the UK I can hardly expect blazing speed from California :)

 

May be something to look into. Perhaps a different upload status indicator would be worth installing, or giving people a way to disable it entirely for situations like this. Also, I didn't think to screenshot what I was up to. I'll try it again with a different file later and see what happens.

[Arthmoor]

Nevermind on what I thought was a workaround. I apparently just got lucky with the Feldscar update the other night. I tried the same trick with the UL Dark Forest update and still ended up with the same problems.

[Yossarian22]

Hi

 

I just want to let you know, that since a while everything is running perfectly fine.

I never got any interrupt again no matter the file size and also the upload rates

are much faster now.

 

Whatever you did to make it work: :thanks: