http://www.digg.com/apple/Handbrake_now ... 0_for_iPod
This was not expected, and it hurt a bit - I'm still working to reshuffle network resources to soak up the sudden bandwidth spike, but I think we'll weather the storm.
Rodney
We got Dugg
Forum rules
*******************************
Please be aware we are now using GitHub for issue tracking and feature requests.
- This section of the forum is now closed to new topics.
*******************************
*******************************
Please be aware we are now using GitHub for issue tracking and feature requests.
- This section of the forum is now closed to new topics.
*******************************
We got Dugg
Last edited by rhester on Thu Jan 04, 2007 11:43 pm, edited 2 times in total.
I am slowly beginning to recover from the Digg storm. I have shunted the download traffic off to a dedicated vhost on my machine where I can throttle the bandwidth.
Things of note:
- We were Dugg. This would ordinarily be complimentary, but at this point, it only served to nearly completely knock us out of service for hours with no benefit to anyone.
- Once we get to a public release, we are clearly going to need _FAR_ more bandwidth than residential DSL can offer. =)
- As a result of the Digg, the public appears to be of the impression that the latest test binaries are public, official releases. This was the very last thing that was wanted. Had I any idea this would occur, I would never have advocated or supported distribution of test binaries beyond the developers themselves.
So...while I can and will continue to support the site from a web-infrastructure standpoint (forums, blog, svn, and Trac), I cannot and will not support this sort of gargantuan mass download scenario. I just don't have the resources.
Suggestions?
Rodney
Things of note:
- We were Dugg. This would ordinarily be complimentary, but at this point, it only served to nearly completely knock us out of service for hours with no benefit to anyone.
- Once we get to a public release, we are clearly going to need _FAR_ more bandwidth than residential DSL can offer. =)
- As a result of the Digg, the public appears to be of the impression that the latest test binaries are public, official releases. This was the very last thing that was wanted. Had I any idea this would occur, I would never have advocated or supported distribution of test binaries beyond the developers themselves.
So...while I can and will continue to support the site from a web-infrastructure standpoint (forums, blog, svn, and Trac), I cannot and will not support this sort of gargantuan mass download scenario. I just don't have the resources.
Suggestions?
Rodney
I think the obvious answer to your distribution problem is bittorrent.
As long as you have enough Good People out there willing to keep "seeding" after they've finished their downloads, the only load on your server would be the tracker and people downloading the torrent file itself. (And you could put the file on an off-site file host.)
If you're not familiar with bittorrent, Wiki it. It's pretty much the best way to distribute a file to a large mass of people while requiring the least bandwidth from your poor DSL line.
(A great Mac app for bittorrent is Azureus.)
Loving the latest beta. Not sold on the interface (too much extra clicking for set up basic things like subtitles and de-interlace), but it does seem to work more reliably than the original 0.7.1. Keep up the great work, guys!
As long as you have enough Good People out there willing to keep "seeding" after they've finished their downloads, the only load on your server would be the tracker and people downloading the torrent file itself. (And you could put the file on an off-site file host.)
If you're not familiar with bittorrent, Wiki it. It's pretty much the best way to distribute a file to a large mass of people while requiring the least bandwidth from your poor DSL line.
(A great Mac app for bittorrent is Azureus.)
Loving the latest beta. Not sold on the interface (too much extra clicking for set up basic things like subtitles and de-interlace), but it does seem to work more reliably than the original 0.7.1. Keep up the great work, guys!
Ok, can we stick a .htaccess / .htpasswd files up so that only us Developers can access the binaries, just distribute the username and password between us and don't give it out to any people on digg 
Then i suggest that if we do have a build we want to make public one of us can host it at another site with a fast backbone
Then i suggest that if we do have a build we want to make public one of us can host it at another site with a fast backbone
Speaking as a non-dev, I am more than happy to put up with the quirks of alpha-phase software when it means a slight improvement over what we had before, and as long as your bugs don't melt my G5 down to scrap-iron, we're cool.dynaflash wrote:In fact, regarding the binaries, I think we should consider if we want that many binaries out there that have an "a" in the version if you know what I mean... Just food for thought...
If a revision comes along that is a little too problematic, I'll just step back a release or two. I think most people in this day and age know what they are getting when it comes to pre-release software.
Besides, if you constantly wait for perfection before letting binaries out, your app will NEVER get used. Mac the Ripper 3.0 has been in beta for about a year now, with no end in sight. VLC has been a superior to Apple's QuickTime Player in many ways for years now, and they still have not gone gold with a 1.0 release.
It's the nature of small-scale software projects, especially free or "shareware" ones, that you pretty much never get out of beta. As long as the app does what people need it to, and continues to improve, most people will be cool with that.
Okay, as SR55 said, we need to limit or take down the binaries for right now. As you said, rhester, they were just intended for internal testing. No one wants to go public more than I, but we also have to be able to keep our dev site up.
1. limit access to binaries or take them down. I vote for limit.
2. Definitely lock down the svn only to those with login and/or password. If the binaries are gone, savvy folk will start hitting the anonymous svn for sure, plus we were going to do that anyway!
3. Get setup with our website (like you started to do before this) so we can shuffle people there to get them up to speed with what is going on and where we are. Information is key here, once people know what is going on, I think most will understand and be fairly patient (at least for a little while) if they know what is coming.
4. Start figuring out what we have, and if it should be public,then get that infrastructure together with the bit of breathing room 1-3 should buy us.
Summary, we have come a very long way in a VERY short period of time. I personally think that feature wise we are pretty close to at least a 0.7.2 public. However, there is alot of detail at least in my mind, to sort out and come together on. Also, wouldnt mind some more stability testing before public. It can be very tough to put the software genie back in the bottle. Something out with a major flaw (ie. ScanController caused crash with just the right scenario in past rev.) and we will get inundated with support stuff on the forums. While that is always unavoidable, I dont think we want to invite any more than we have to.
Just my .02.
1. limit access to binaries or take them down. I vote for limit.
2. Definitely lock down the svn only to those with login and/or password. If the binaries are gone, savvy folk will start hitting the anonymous svn for sure, plus we were going to do that anyway!
3. Get setup with our website (like you started to do before this) so we can shuffle people there to get them up to speed with what is going on and where we are. Information is key here, once people know what is going on, I think most will understand and be fairly patient (at least for a little while) if they know what is coming.
4. Start figuring out what we have, and if it should be public,then get that infrastructure together with the bit of breathing room 1-3 should buy us.
Summary, we have come a very long way in a VERY short period of time. I personally think that feature wise we are pretty close to at least a 0.7.2 public. However, there is alot of detail at least in my mind, to sort out and come together on. Also, wouldnt mind some more stability testing before public. It can be very tough to put the software genie back in the bottle. Something out with a major flaw (ie. ScanController caused crash with just the right scenario in past rev.) and we will get inundated with support stuff on the forums. While that is always unavoidable, I dont think we want to invite any more than we have to.
Just my .02.
I'm controlling the download traffic now so it doesn't affect the rest of the site, that's not the issue.
The issue is that we got Dugg, and we shouldn't have, because this release (hell, it isn't even a "release" - it's a mishmash of various binaries from various contributors, not necessarily even representing the same svn revision!) wasn't ready for prime-time and we knew it.
This isn't a matter of "looking for perfection" - there were certain _supportability_ goals we needed to reach prior to a public release. Noticed the auto-update functionality is broken? No? We did. And it requires building new infrastructure to make it work again. Many small issues like these were all being worked on when this came out of nowhere.
I appreciate people who are fans of the software. I'm thankful for people who appreciate the small band of developers who came together here to resurrect a project that otherwise would almost certainly have died silently. But we now have a horde (and I do mean in the thousands) of people who all have an unsupportable, known-broken "release" of one form or another (or one _platform_ or another), who will ultimately be complaining of the problems we already know about and seeking relief...which would not have been necessary had an impatient end-user who can read development forums not jumped the gun and "spilled the beans" as if there was some sort of global conspiracy afoot to keep the world away from "HandBrake Gold".
Now, instead of devoting my time to continuing to build out a support infrastructure that can actually be of value to end-users, I have to spend my time trying to protect myself _from_ them. This most definitely wasn't what I got involved for, and I'm more than a little [Censored] off that by linking _directly_ to a binary in Digg the post author managed to:
a) Give the illusion that this is a "real", supported release
b) Give the illusion it came from a m0k developer
c) Completely bypass any need to read our FAQ that quite clearly lays this all out, gives no reference at all to these forums, and puts both us and the "users" in an untenable position at the same time
Why? Because someone wanted their 5 minutes of fame on digg.com. And Lord knows, the post _is_ popular...I know that better than anyone. But worth it? I doubt it.
The only reason I haven't completely disabled downloads at this point is because the poor downloaders aren't at fault here. They had no idea they'd been duped, and I don't think it right to replace their binaries with a "ha - gotcha!" message inside a tiny ZIP file.
So, all I have to say is the following at this point:
- We really need to revisit how we distribute test binaries - or, even better, whether or not we even should now that all platforms are svn-compilable ready
- Where we're going to come up with a commercial-grade web server for static downloads once we really do release...and take my word for it, not even a dedicated T1 will be enough. I'd consider starting with a fractional T3, but who's going to pay for it? T1s were enough during the dial-up days, but no longer in the modern world of asymmetric high-speed connectivity.
- To kenthorvath on Digg: I honestly don't much care whether your intentions were noble. If you knew where to find the binaries, you had to do a bit of reading to get there, and should therefore have known better to do something so blindly stupid. Consider yourself thankful I actually worked to recover this site rather than just stop apached once and for all and say to hell with all of it. I have better things to do with my time than fight off unintentional DoS attacks.
Sorry for the poor attitude, guys...I'm just beyond irritated at the moment.
Rodney
The issue is that we got Dugg, and we shouldn't have, because this release (hell, it isn't even a "release" - it's a mishmash of various binaries from various contributors, not necessarily even representing the same svn revision!) wasn't ready for prime-time and we knew it.
This isn't a matter of "looking for perfection" - there were certain _supportability_ goals we needed to reach prior to a public release. Noticed the auto-update functionality is broken? No? We did. And it requires building new infrastructure to make it work again. Many small issues like these were all being worked on when this came out of nowhere.
I appreciate people who are fans of the software. I'm thankful for people who appreciate the small band of developers who came together here to resurrect a project that otherwise would almost certainly have died silently. But we now have a horde (and I do mean in the thousands) of people who all have an unsupportable, known-broken "release" of one form or another (or one _platform_ or another), who will ultimately be complaining of the problems we already know about and seeking relief...which would not have been necessary had an impatient end-user who can read development forums not jumped the gun and "spilled the beans" as if there was some sort of global conspiracy afoot to keep the world away from "HandBrake Gold".
Now, instead of devoting my time to continuing to build out a support infrastructure that can actually be of value to end-users, I have to spend my time trying to protect myself _from_ them. This most definitely wasn't what I got involved for, and I'm more than a little [Censored] off that by linking _directly_ to a binary in Digg the post author managed to:
a) Give the illusion that this is a "real", supported release
b) Give the illusion it came from a m0k developer
c) Completely bypass any need to read our FAQ that quite clearly lays this all out, gives no reference at all to these forums, and puts both us and the "users" in an untenable position at the same time
Why? Because someone wanted their 5 minutes of fame on digg.com. And Lord knows, the post _is_ popular...I know that better than anyone. But worth it? I doubt it.
The only reason I haven't completely disabled downloads at this point is because the poor downloaders aren't at fault here. They had no idea they'd been duped, and I don't think it right to replace their binaries with a "ha - gotcha!" message inside a tiny ZIP file.
So, all I have to say is the following at this point:
- We really need to revisit how we distribute test binaries - or, even better, whether or not we even should now that all platforms are svn-compilable ready
- Where we're going to come up with a commercial-grade web server for static downloads once we really do release...and take my word for it, not even a dedicated T1 will be enough. I'd consider starting with a fractional T3, but who's going to pay for it? T1s were enough during the dial-up days, but no longer in the modern world of asymmetric high-speed connectivity.
- To kenthorvath on Digg: I honestly don't much care whether your intentions were noble. If you knew where to find the binaries, you had to do a bit of reading to get there, and should therefore have known better to do something so blindly stupid. Consider yourself thankful I actually worked to recover this site rather than just stop apached once and for all and say to hell with all of it. I have better things to do with my time than fight off unintentional DoS attacks.
Sorry for the poor attitude, guys...I'm just beyond irritated at the moment.
Rodney
Correct on all accounts.
I just checked and they are starting to mirror the downloads. So, my guess is that we will shortly get hit with all sorts of questions and what not on the forums.
I would expect the m0k forum to get nailed as well.
It may be wise, at least to get something up for when they come. .. And I do believe they will come....
I just checked and they are starting to mirror the downloads. So, my guess is that we will shortly get hit with all sorts of questions and what not on the forums.
I would expect the m0k forum to get nailed as well.
It may be wise, at least to get something up for when they come. .. And I do believe they will come....
Bittorrent seems like a good option, you could also temporarily upload the files to http://la.gg to get you by the Digg effect.
LA.GG is your #1 free media host, currently hosting any file types, including:
* Images (.jpg, .gif, .png)
* Videos (.avi, .wmv, .asx, .asf, .mov, .mpg, .mpeg)
* Music (.mp3, .m3u, .wma, .wax, .ogg, .wav)
* Flash (.swf)
that doesn't appear to be much use. Lets worry about releasing it when its ready shall we?
My patience runneth out with the following Digg reply:
by DigitalN 8 minutes ago
well some of us don't care if it is not meant for the public or not, we dont want to have to rip a DVD with handbrake for 2 hours, and then have to use iSquint for another 2-3 just so we can use it on our iPod.
OK, so he doesn't care. Good. Neither do I. Downloads now marked as "gone" via Apache redirect and are no longer available via Web (still maintained on the FTP).
Rodney
by DigitalN 8 minutes ago
well some of us don't care if it is not meant for the public or not, we dont want to have to rip a DVD with handbrake for 2 hours, and then have to use iSquint for another 2-3 just so we can use it on our iPod.
OK, so he doesn't care. Good. Neither do I. Downloads now marked as "gone" via Apache redirect and are no longer available via Web (still maintained on the FTP).
Rodney
