«interrupted: network error» when downloading through a VPN connection and failure to handle ReCaptchas adequately
quipeccavit last edited by
Dear Ladies and Gentlemen!
The «interrupted: network error» problem has been reported often enough by plenty of people, for instance on Quora, RSSing, Sayugi, VPN Crew et al, as well as in several places here in the Opera user forums.
Given how many different people suffered from the browser's shortcoming over such an extended period of time, still experiencing the problem to this day in the current version of Opera (66.0.3515.72 at the time of this writing), there is no denying that the problem is on the browser's end and has nothing to do with the operating system, antivirus, firewall, ad-blockers, extensions, add-ons or other excuses and distractions that administrators and moderators like to use as a pretext for closing/ignoring/locking threads with respective reports prematurely. The fact that the problem is «old» and has still not been resolved makes it worse, not better, so instead of locking the respective thread, it needs to remain open for current and future users to add their observations, until the bug has finally been fixed or sufficient resources/VPN servers lead to its disappearance. It makes much more sense to have one place containing all the reports, than to force people to open another one each time just because moderators try to hide the embarrassment.
Yes, it now takes longer for the network failure to present itself, usually after hundreds of MegaBytes, mere seconds before finishing the download, but that's even worse than before, because more precious time is being wasted waiting in vain. Second, download speeds are often dropping over the course of a download, starting at 100 kB/sec, then slowly falling to 80, 50, 20, 10 kB/sec. Since this results in a long time before downloads finish, the risk of being affected by the dreaded «interrupted: network error» augments further.
To add insult to injury, the forwarding/rerouting of the scourge of the Internet, Google's goddamned Cloudflare ReCaptchas, often takes too long or does not work at all, because the site «could not connect to the ReCaptcha server». Since every idiot and their brother imposes that harassment on their visitors these days, even when/where it really is not necessary, this results in having to click even more stupid stamp-sized blurry pictures with tiny pixels of bicycles, buses and pedestrian crossings. And since this happens so often that one has to request/resolve a lot of that nerve-killing impertinence, the occasions where «your computer/network requests a lot of ReCaptchas, try again at a later time» become more frequent as well, prohibiting access to the download or website for a while.
All these problems seem to indicate if not a bug, then an insufficient quantity/structure of servers and bandwidth; the fact that they have not been resolved in years justify the suspicion that not enough servers and bandwidth might be to blame, rather than a bug that's being kept as a pet instead of being squished.
This is something that only insiders can tell for sure, but merely ignoring, blocking, closing and locking reports or blaming add-ons and extensions is no acceptable course of action and definitely no solution. After years of dealing with these frustrations, it is high time to finally take them serious and do something definitive about them.
Thank you for your attention, keep up the otherwise good work and have a pleasant weekend.
blackbird71 last edited by blackbird71
@quipeccavit Although I don't have a horse in the VPN race (I never use it), there are some factors to keep in mind. First, many site and relay servers keep track of dead time on a connection and drop it if that times out, which in turn will cause a variety of error messages, depending on what other element(s) in the pathway get hiccuped by the drop out. Dead time can be caused by any number of things: a clogged relaying server anywhere in flow, high packet latency anywhere in flow, and slow packet processing by software anywhere in flow (including at each connection end - the user and the target website). Since there are so many elements involved (including potentially-different timer settings at every server and packet-touching software in the chain), there will be a lot of "your mileage may vary" effects that can be observed, which is why user software (AVs, extensions, OSs, etc) are often cited as contributory.
Second, bandwidth limitations in a VPN's server farm certainly can be a major factor, since it can by itself cause a connection drop and can also act to greatly increase packet latency even if not directly dropping a connection. What must be kept in mind is that VPN servers and their bandwidth cost somebody something, somewhere. If the primary service objective of the VPN is to provide users a basic private browsing service, it may not be cost-effective for the VPN operator to invest in expansion of servers and bandwidth in order to meet the needs of the occasional large-file download user. Since I don't work for Opera, I certainly don't know their official view on this, but my instincts indicate it is probably a significant factor, since both the browser and the VPN feature are provided "free".
quipeccavit last edited by
That's all good and fine and correct and nothing new or unexpected, but no reason not to remedy a problem that has been bugging users since years.
As for the financial situation, Opera's revenues grew over 100% since 2018, where they made a net income of $35 mio. USD. The annual report for 2019 is not going to be published until February 25th, 2020, but ceteris paribus (lat. = all other influencing variables and factors remaining equal), that growth in revenue should be sufficient to get a grip on the annoying failures of downloads and slow download speeds via VPN.
Yes, other companies serve their shareholders first and customers last, but Scandinavians are a cosmopolitan, open-minded and friendly kind of people with a social conscience and honest business morale nearly free of corruption. The boys from Norway have always been better than other tech companies, so here is hoping that despite the acquisition by Chinese capitalists, a little bit of that spirit remains. Besides, offering a product that works reliably and ironing out its flaws, shortcomings and annoyances is what ensures that a company's growth continues in the future, whereas maximising shareholder value is a myopic strategy of managers/owners who are preparing to cash out.
Look, I wouldn't have bothered to bring this up if it had been a couple weeks and if the inconvenience was due to a bug that's being worked on. After years of broken or slowly crawling downloads, however, it is time to do something about it and make sure that such a basic functionality of any web browser finally functions properly and reliably, is it not?
quipeccavit last edited by
Is it just me or has Opera decided to implement a peculiar 'workaround' to bully its users with interrupted network errors now?
In the current version 67..., I get slightly fewer «interrupted: network error» problems (still around 10%, though), but downloads are now even slower than they were before, beginning at between 50-100 kB/sec and then gradually declining to 20 or even 10 kB/sec. On larger files, this is devastating.
Furthermore, despite having the flag for parallel downloads enabled, the stupid software slows down one file to 0 kB/sec, 'speeds up' the second download to 20/50/100 kB/sec for a few seconds and then does it the other way round, i.e. increasing the speed on the first file why the second one is briefly 'parked' at 0 kB/sec. This idiotic behaviour of increasing and decreasing speeds between the files that are being downloaded results in endless waiting times until a file has finally finished downloading, unless the dreaded interrupted network error gets the better of it during the endless process of downloading.
Firefox, TOR, Vivaldi and a bunch of other browsers do not do this, not even natively with their own means. It goes without saying that there are tons of download managers out there that handle the basic functionality of file downloads more intelligently, too, so what gives? Can the basic functionality of file downloads not be implemented in a half-decent manner?