Sarah Passerine Posted July 15, 2013 Share Posted July 15, 2013 Ever since SSB went live on LeTigre regions, no one is baking for me. However, it only happens when I am using a cellphone connection. Is there some issue with the robustness of the code that is making the SSB stuff fail for certain connection types/routes/whatever? Link to comment Share on other sites More sharing options...
Darien Caldwell Posted July 16, 2013 Share Posted July 16, 2013 SSB is only on about 1/20 of the Grid. so very unlikely it has anything to do with your issues. Link to comment Share on other sites More sharing options...
Sarah Passerine Posted July 16, 2013 Share Posted July 16, 2013 Not so. If I log in to a release channel region, things work. My sim, Dolly Dreams, is on LeTigre, and is using SSB. Avatars are not baked properly for me. If I go to Myanimo, another LeTigre sim, the same thing happens. I think it is something that is going on with the combination of SSB and T-Mobile internet. I'll post the debug console messages tomorrow, if anyone cares. Paraphrasing, there are some odd messages about a host not being found. Link to comment Share on other sites More sharing options...
Whirly Fizzle Posted July 16, 2013 Share Posted July 16, 2013 Heya Sarah, You should file a bug report on the JIRA about this. Run a session on the latest release version of Viewer 3 while connected to your T-Mobile network. Login to a non-SSA region where there are other avatars, wait 10 minutes and then teleport to an SSA region where there are other avatars, stay there for 10 minutes or so and then teleport again to a non-SSA region where there are avatars and stay there for 10 minutes or so and then logout. Before relogging, grab your logs folder, zip up the logs folder and attach it to your JIRA issue using More Actions -> Attach files. Instructions on where to find your logs folder can be found HERE. Repeat the above process running a session on the same viewer with the same avatar while connected to your normal network which does not have these problems & also attach your logs from that session to the JIRA issue. Link to comment Share on other sites More sharing options...
Sarah Passerine Posted July 16, 2013 Share Posted July 16, 2013 Alright, will start putting that together. Edited to add: The gist of it is there are a lot of CURL GET FAILED and HTTP GET failed messages, with the status code of 00000012 and reason of "Transferred a partial file". Link to comment Share on other sites More sharing options...
Lindens Monty Linden Posted July 16, 2013 Lindens Share Posted July 16, 2013 Sarah Passerine wrote: The gist of it is there are a lot of CURL GET FAILED and HTTP GET failed messages, with the status code of 00000012 and reason of "Transferred a partial file". Hmmm, that will be interesting to see, so please do file a Jira and report the number here as well. That suggests incomplete data from servers. I'd ask for one additional thing: when you create the jira, include a SecondLife.log file from as recent a session as possible. Link to comment Share on other sites More sharing options...
penfold83 Posted July 16, 2013 Share Posted July 16, 2013 Sarah, the t-mobile net you have is it 3g ie from a dongle? If so whats your respetion like does it drop at all ? just wondering if that was the prob with it. Link to comment Share on other sites More sharing options...
Sarah Passerine Posted July 16, 2013 Share Posted July 16, 2013 Logs uploaded at the bug filed at: https://jira.secondlife.com/browse/BUG-3323 Please forgive if I did something wrong, this was my first report! Also, penfold83, the connection was a steady HSPA+ 21.6 Mb/s link. When it is spotty, SL gets incredibly weird, with messages posting out of order, and multiple times too! In this case, something else is going on. Link to comment Share on other sites More sharing options...
Lindens Monty Linden Posted July 16, 2013 Lindens Share Posted July 16, 2013 Looks fine, thanks. Link to comment Share on other sites More sharing options...
Lindens Monty Linden Posted July 16, 2013 Lindens Share Posted July 16, 2013 FYI, the following error: CURL GET FAILED, status: 00000012 reason: Transferred a partial file indicates that short data was returned during an HTTP response. I.e., viewer had a reason to receive 'X' bytes but received less at the end of the response (usually a connection close). This points to a problem upstream of the viewer: nic, local switch, local router, carrier, servers. Link to comment Share on other sites More sharing options...
Sarah Passerine Posted July 16, 2013 Share Posted July 16, 2013 If it is some sort of unfixable problem due to the network type, then at least we know the new method isn't as robust as the previous way of doing things. Link to comment Share on other sites More sharing options...
Lindens Monty Linden Posted July 16, 2013 Lindens Share Posted July 16, 2013 Well, we'll have to see where the errors are generated first. But if it is a NIC/adapter/carrier issue, it means this is likely happening all the time for you. Software is either unable to detect the problem and ignores it or detects it and retries. In the first case, you may get data corruption, in the second case, wasted capacity. Link to comment Share on other sites More sharing options...
Darien Caldwell Posted July 17, 2013 Share Posted July 17, 2013 Sarah Passerine wrote: If it is some sort of unfixable problem due to the network type, then at least we know the new method isn't as robust as the previous way of doing things. Except that the failure could be completely coincidental, and have nothing to do with SSB. If you stub your toe the same day Justin Bieber releases an Album, you can't with certainty correlate that the Album release caused your toe to stub.. Link to comment Share on other sites More sharing options...
Sarah Passerine Posted July 17, 2013 Share Posted July 17, 2013 This is why the conversation has moved to the jira. The issue has been traced to the viewer receiving bad Content-Length headers from http://bake-texture.agni.lindenlab.com. I don't know if this is happening on LL's side, T-Mobile's side, the NSA's side, or whatever, but at least the problem has been identified! Link to comment Share on other sites More sharing options...
Darien Caldwell Posted July 18, 2013 Share Posted July 18, 2013 Sarah Passerine wrote: This is why the conversation has moved to the jira. The issue has been traced to the viewer receiving bad Content-Length headers from http://bake-texture.agni.lindenlab.com. I don't know if this is happening on LL's side, T-Mobile's side, the NSA's side, or whatever, but at least the problem has been identified! My bet would be on T-Mobile. Mobile internet providers are known for doing a lot of tampering with HTTP traffic. In fact there are even documented cases where they decrypt HTTPS traffic to HTTP and then spoof it as still being HTTPS to speed up transfer times. As a general rule, Mobile internet isn't very trustworthy, reliable, or robust. Link to comment Share on other sites More sharing options...
Bunderwahl Guisse Posted August 6, 2013 Share Posted August 6, 2013 I have the same issue but with metro pcs as the provider. The annoying thing was that before the roll out everything was great, changing, voice, streaming media. Multi tasking on the net. Since everyone is gray except prims and mesh (which rezzes instantly). Others see me fine, I see them and myself as a gray manikin Link to comment Share on other sites More sharing options...
tamaith Aeon Posted August 21, 2013 Share Posted August 21, 2013 I am having the same issue on Verizon's wireless network. Glad I am not the only one and that I know what the problem is now. I was tearing my mesh hair out over this issue. Link to comment Share on other sites More sharing options...
Sarah Passerine Posted August 22, 2013 Share Posted August 22, 2013 Well, it's nice to know it isn't T-Mobile specific. Perhaps it is more likely to be fixed. Link to comment Share on other sites More sharing options...
Sarah Passerine Posted August 22, 2013 Share Posted August 22, 2013 Also, oddly enough, baking was working on LeTigre for a week, until yesterday. Link to comment Share on other sites More sharing options...
Elyse Sirnah Posted August 22, 2013 Share Posted August 22, 2013 I'm having the same problem, but I am on Verizon.. funny thing is I can use my phone as a hotspot and see everyone just fine and dandy, if I use my hotspot (same network and still has unlmited data) EVERYONE is gray!! UGH! Link to comment Share on other sites More sharing options...
Bunderwahl Guisse Posted August 22, 2013 Share Posted August 22, 2013 I am connecting onthe phones 4G with PdaNet..via usb cable. Anyone else using that as well? Link to comment Share on other sites More sharing options...
Pinkcat Carter Posted August 25, 2013 Share Posted August 25, 2013 same problem in europe too. avas are complete grey Link to comment Share on other sites More sharing options...
Bunderwahl Guisse Posted August 29, 2013 Share Posted August 29, 2013 Any news on this yet? LL appears aware do not know how easy a fix it would be, seems like just a setting change, or having the server not check, or possibly being able to force a bake. There were two days when I could see again, not sure what happened, thought there had been a fix. At this point I am starting to think of inwordlz as a substitute Link to comment Share on other sites More sharing options...
Lindens Monty Linden Posted August 30, 2013 Lindens Share Posted August 30, 2013 I'm assuming that all the cases here involve laptops tethered to cellphones. Are there any other configurations here? Wireless broadband via USB dongle? Some other Frankenstein technology? As an experiment, would those experiencing problems set the following two Debug Settings before logging in and then re-test: TextureFetchConcurrency to 2 MeshMaxConcurrentRequests to 2 You get to these by: Enabling debug menu with Ctrl-Alt-D Select the 'Debug > Show Debug Settings' menu item Type a bit of the name or scroll through the list to find the settings After the test, you can reset the settings to their original values. Otherwise, you'll have a very, very slow scene build when you go to new regions. Link to comment Share on other sites More sharing options...
Bunderwahl Guisse Posted August 30, 2013 Share Posted August 30, 2013 I have unlimited data on a MetroPCS phone (believe they are owned by T mobile) and use a app called PdaNet to connect my desktop to the internet via a usb cable attached to the phone. The connection allows me to DJ in sl, watch streaming movies and the rest of SL rezzes quickly with no issues other then clothing/tattoo/skin layers on avatars not rezzing. All worked perfectly before SSA I tried the debug settings suggested, no change Link to comment Share on other sites More sharing options...
Facebook
Instagram
Twitter
YouTube
Flickr
Recommended Posts
Please take a moment to consider if this thread is worth bumping.
Please sign in to comment
You will be able to leave a comment after signing in
Sign In Now