|
|
|
[
Permlink
| « Hide
]
Marion Rickenbacker added a comment - 28/Mar/07 04:47 AM
Is anybody addressing these log in issues? It is extremely frustrating that apparently trivial issues and new features get addressed quickly, while the basic inability to access Second Life is not
I have been having this same problem and it causes my computer to slow down for a few minutes and I have tried emptyiing TMP didnt help, using a wired connectiong, didnt help, and re-downloading, that didnt help.
This is a presence issue – it sounds like you recently logged off, but the system didn't get the memo.
We need some more detail here – does this happen every time? Did you just log off? Is there any pattern to be seen? etc I have not had the problem of "your service will be available in (5 minutes from the current time)", but I have consistently had the "Connecting to Region" hang. (49 times out of 50, perhaps, with 1.15)
The last message seen on the console was always: 2007-04-26T11:46:04Z INFO: LLInventoryModel::buildParentChildMap() 2007-04-26T11:46:04Z INFO: LLCircuit::addCircuitData for 66.150.244.151:12036 though in 1.15 it is sometimes: (I believe this message appears /after/ the addCircuitData message) This has happened just as frequently in 1.13, 1.14, and (seemingly more frequently) in 1.15. In 1.13, I was able to work around the problem by connecting with libsecondlife's TestClient, (with which I have never witnessed any similar issue). Without fail, successful connection with the TestClient resulted in my next attempt at connecting with the official client being successful. After the forced upgrade to 1.15, this trick no longer works: connecting with the TestClient does not change my inability to connect with the official client at all. When hanging at "Connecting to Region", CPU usage drops to zero, and I have waited ten minutes with no response. The secondlife window also stops repainting, responds to no clicks, and needs to be kill'd (ctrl-z, kill %; kill %; works. Note that it needs two kills to end.) Usually upon killing secondlife and returning, I receive the message "region has begun the logout process, please try connecting again in a few minutes", and am able to re-attempt connection immediately after receiving such a message. If roughly ten minutes pass between attempts, I do not receive that message, though usually am still unable to get passed "Connecting to Region" The vast majority of my connection attempts end in the region hang. I believe that after not being connected for several hours, every initial attempt has failed in this way (again, in versions 1.13, 1.14[ both ], and 1.15[ 1.15 seeming to be much worse ]) I have attempted reinstalling secondlife, I have no firewall blocking outgoing ports (and have seen no required incoming ports listed), and this occurs regardless of what region I am connecting to, time of day, etc. I am directly connected to the modem, I do not go through a separate gateway. Debian Linux 2.6.18-4-686 #1 SMP Mon Mar 26 17:17:36 UTC 2007 i686 GNU/Linux Subversive,
Can you remove your router from your internet connection, plug your computer in directly (configuring the IP address to what your router's was), and see if the issue still exists? Oh, you might have already done that. Does your modem do NAT?
bushing's comment is exactly right. There's a lot of bugs that can cause this problem, and the devil is in the details. The problem described here happens when something (client crash, sun storm, evil gnomes, whatever) causes your presence info to get scrambled on the server. We detect that there was a problem, lock you out until the system can figure out what's going on, and then let you back in. So, the fact that you get locked out doesn't tell us anything. What happened before you got locked out makes all of the difference.
in order:
1) I am connected directly to the modem, and always have been 2) This means there is no NAT involved - I do in fact have a real IP address and use dyndns to connect to my home system often. There's no Router, no NAT. 3) Earlier today, a friend came over and I used his (window-xp) Laptop to connect (using the system which has this issue as a gateway). It had no trouble connecting. I simultaneously attempted connection on the system which has been having the issue, and was unable to connect. This appears to be an issue specific to the Linux client. 4) "your service will be available in five minutes", it's obvious that this means my connection status is out of sync, but this only happens after many repeated attempts. [perhaps this is a separate issue?] The majority of the time, I do not get the "..[five minutes from now]" message. The vast majority of the time, there is nothing which happened "before" this, it just happens. For example, I just got home, tried to connect, and hung at "connecting to region". Tried again, same thing. (wash, rinse, repeat). Instabilities with an Alpha release being what they are, of course sometimes things do occur before I see this issue, but "connecting to region.." hangs have remained a constant as other things come and go. And again, I have never had this issue with the Libsl Test Client (which I occasionally confirm). I am not often running libsl-related programs, and had this issue before ever downloading libsl. I sometimes use it as a backup when the client refuses to connect [or at work ;)], and have never seen the issue. This is all on the same system. I am willing to do whatever you want in order to help resolve this issue, I just need to know what you need. In case it wasn't clear, whenever I have said "unable to connect" in the previous comments, I am referring to the "connecting to region" hang.
Well Rob, the thing is this can happen at any time lol. For me it is always following a crash, and in 1.14/1.15 that can now be doing absolutely anything, including opening profiles, notecards, standing still, doing nothing, clicking on something, laggy sim, clicking a window, typing an IM, opening a window, turning around, changing mode or being quit by the system eg region going down... well you get the picture.
What then happens, normally when I relog, I get a failed region starting login process, which is usually erroneous in any case as other people report no sim down or crash, so the system or my client quits me again. Then when I try to relog, I get the wait 5 minutes. I understand the reason for 5 mins is probably anti griefer/hacker, but it should allow more than two, especially when one is an errpneous system quit. Further, this is happening more frequently with 1.15, and prior to 1.15, you could often get around the 5 mins by somtimes quitting, other times clearing network cache, sometimes hard drive cache too, or changing login destination from last loc to home, but this no longer seems to work. Very irritating if you are in the middle of a meeting, more so if it's the middle of an amorous moment I forgot to mention: iMac G5, 1Gb RAM, OS X 10.4.9, ATI Radeon X600 Pro.
I am getting a lot of extended connecting times and what appear to be freezes, but this may be unconnected, I don't usually get the 5 min notice when force quitting from that situation. Experiencing too many lock out, even after a CLEAN quit !
Even cleaning cache does not help SL should clean all previous data when a new auth is done... SL the service or SL the viewer?
A fresh install with a deleted ~/.secondlife directory has the same issue. Raising Bug level as it remove the ability to connect for most of the day.
30+ consecutive login tries and no region connet ! Things gone worse with version 15, but I think this is not related to the client. This problem already happened with version 14, just before the .15 release. Something changed on servers side so. Seriously, anything you guys want me to do: Need me to use a debug build? Run it through gdb? Give you complete output from Ethereal? Want to set up a secure remote connection to my system? Anything you want.
Running SL through strace seems to make login possible. (I never had been able to see where's the lock because it runs OK everytime I try to debug).
Please confirm by lauching: strace -f secondlife just freezes my system before I get a chance to try logging in (though not before lots of info is displayed)
My suspicion has been that this involves some race condition, and though I haven't been able to confirm your strace experience, it would seem that yours adds further evidence to such.
I have attempted banging on the keyboard a bit just before the "addCircuitData" message, and, placebo or not, I am now logged in. This is happening more frequently. Kuula sim was full of "ghosts" the other day and once able to relog in those people reported that they had crashed then continued to get the "region is currently logging you out..." message.
Same happened to me today, I relogged as the result of a persistent hourglass when trying to open my inventory. I was still able to move etc. Upon relogging I got the "region is logging you out..." and it persisted for over 30 minutes. Sent a friend to that sim and she bulldozed me into the neighboring sim, within 5 minutes I was able to log back in. Subversive, try logging into a different region when this happens.
If this is just a presence issue, it is normal for the connection to temporarily freeze or even tell you that you have to wait and try again. Otherwise, the description does not explain enough of what kind of "freeze" happens.
As previously mentioned, this bug occurs regardless of what region I am attempting to connect to. I suppose I should clarify that I mean by this:
Also perhaps notable: This is ONLY a login issue. Once connected, no similar issues occur. This is most certainly not a presence issue, either. This happens during the "first connection of the day", often. There seems to be some confusion in this bug, as the original description mentioned both the "Connecting to Region" hang /and/ the ",,,[five minutes from now].." message which sometimes occurs after a disconnection. Half the comments seem to be talking about the "five minutes from now" issue, which, in my experience, only very rarely occurs. I plead again: Please tell me whatever additional information may help resolve this, instead of just saying "well, it works for /me/..." Since you're on Debian: try the command 'traceroute' or 'ping' to the address after addCircuitData. Like:
$ traceroute 66.150.244.151 Note the output Gone with 15.1.3
Back with 16.0.5 pisses me off PING 66.150.244.151 (66.150.244.151) 56(84) bytes of data. tracepath 66.150.244.151 I have not had this issue in a few days now, though I have had it since the most recent client upgrade at least twice.
I was able to connect yesterday to SL on both my laptop and my home PC. I used my home PC to connect to SL this morning at approx 0230 PDT from my PC, worked fine. I am now out of state 12+ hours later. When I try to login to SL I get the following error "Unable to login. Login Packet never received by login server. Please try again in a few minutes. Or click help for advice and a link to the system status page". I have been trying to login for over an hour, I uninstalled SL version 1-16-0-6. Downloaded and re-installed. Same error, I even tried connecting to a different region. Same error. I have never had this issue before.
As expected, I spoke too soon. Same hang on add circuit data two days in a row now. Back to where to were.
Is it possible to delete Addy Babii's comment, as it is describing an utterly unrelated issue? This issue is confused enough. I'm having the same problem on my AntiChrist Dagger account, but my secondary can log in fine. Every single time i try to connect as Anti, the bar stops at Connecting to Region and needs to be closed in the Task Manager. Since it started 3 days ago i havent been able to log onto AntiChrist
This is apparently a server capacity or link capacity problem.
From Europe, link losses rises when going towards 18.00 CEST, then is totally impossible to connect after 19.00 CEST up to 1.00 CEST. This matches evening connections for europe and end of work day connections for USA... This bug is really annoying, but look at the currently assigned bugs and the total of bugs: nobody cares. There's even inventory losses bugs and they don't assign enough people to get rid of all those critical bugs. They prefer creating new ones with new more beta features. Again, as far as I can tell this ONLY effects me on Linux. While it may be related to server load, I have side-by-side been able to connect on Windows while not being able to connect on the Linux client. Always freezes at "Connecting to Region.."
I am also experiencing this bug. While it may be server-related, I think there is definitely a client issue here. When the freeze occurs, the client seems to stop responding to X expose and resize events, which in my opinion indicates a mutex issue of some sort.
I'm running the client (1.17.0.12) in gdb right now, and am at the point where the freeze has occured. I've borken and continued several times just to be sure, but a stack trace indicates that the freeze is occuring in libfmod. Program received signal SIGINT, Interrupt. It looks to me like fmod is blocking, waiting for something that never happens. Considering that ALSA works just fine with all of my other ALSA-enabled applications, I am inclined to guess that the bug is in fmod, rather than the kernel drivers. My guess is that addCircuitData is working just fine, and the client is moving on to something sound related, but crashing before it gets a chance to spam anymore messages to the console. Now I'm going to try running with sound off.... And now it connects just fine. Once. Twice. Thrice. Now I'm going to load the OSS modules and disable SL's use of ALSA (which I was forcing it to use before). And it connects fine. Once. Twice. Thrice. Now I'm going to reenable ALSA, just to see what happens. ... And it goes Boom. Now I'm reconfiguring ALSA to not use dmixed to mix sources from multiple processes via software. ... And it still goes Boom. Let's see if installing the ALSA-OSS emulation modules and forcing SL to use OSS, while still having dmixed turned on in ALSA lets us play from multiple processes whilst circumventing the SL lock up... Hm, it doesn't work. It seems to initialize fine, but I get no sound from SL using ALSA-OSS emulation. I'm going to reboot in order to unload the oss modules that don't want to unload and try again, in case they are conflicting. Here's some system information, if it helps with the debugging process: Ok, I got multiple sources and ALSA-OSS emulation working, and I can still log in reliably.
The sound is a little choppy, but here's the workaround (if you still want to use ALSA) that works for me: 1). Edit the secondlife shell script and make sure the following lines look like this: 2). Add the aoss script to the long line near the bottom of the script: 3). Here's my .asoundrc, just in case someone wants it (I don't know how it works; Google gave it to me, lol): #one called "dsnooped" for capturing #and this is the real magic #a quick plug plugin for above device to do the converting magic. saves #a ctl device to keep xmms happy #here we try to point the aoss script to our asymed device ctl.mixer0 { #this sets the default device Not tried it yet, but I just wanted to comment that I've had numerous (I thought unrelated) sound issues. I have tried "disabling" sound because of them, but muting sound in the SL preferences does not appear to actually prevent SL from using any sound drivers whatsoever. (on the other hand, I could be completely wrong).
Still, I wonder what you mean by "with sound off." How do /you/ disable sound in SL? I usually have sound muted due to my other issues, and afaik have still had the connection problems while sound is muted. Again, I could be wrong. Not home at the moment, so haven't tried. I exported all of the BAD_XXX=x variables in the secondlife shell script to turn off sound.
Was having the same issue in last about 2 hours, following new version 1.17.1.0.
At first, I thought that uncommenting all LL_BAD_ESD, LL_BAD_OSS and LL_BAD_ALSA in ./secondlife script, was fixing it, but when I have commented them again, I was still able to connect. [see end of comment, now I believe this had indeed fixed it] So I begin to think this have more to do with ping and traceroute. — 66.150.244.151 ping statistics — paul@Arcturus:~$ traceroute 66.150.244.151 While having the problem, I was trying with 1.17.0.12, trying to log on different regions with Preferences/General Tab/Start Location but was having always stopping on: LLCircuit::addCircuitData for 66.150.244.151:12036 Oh, I have the problem back right now: — 66.150.244.151 ping statistics — — 66.150.244.151 ping statistics —
So, I begin to accept the idea that it has to do with FMOD initialization. Sorry for long comment, guess next time, I'll add in a text file and upload it. Configuration: Second Life 1.17.1 (0) Jun 25 2007 10:55:27 You are at 259984.8, 287258.7, 601.2 in Arsheba2 located at sim3077.agni.lindenlab.com (64.129.44.96:13005) CPU: Can't get terse CPU information [2.66GHz, Celeron 331+ D on a P5GZ-MX motherboard] I am almost sure we do not use esd anymore on Ubuntu 7.04 (feisty).
Read: https://wiki.ubuntu.com/PulseAudio 
I can't find esd on my box, but libpulse0 is there. Guess instructions for Ubuntu 7.04 should suggest to install pulseaudio-esound-compat package. After adding pulseaudio-esound-compat package, I get a new warning:
hehe, I see I have: 2007-06-26T07:28:43Z WARNING: sendMessage - Trying to send RequestImage on unknown circuit 64.129.44.96:13005 2007-06-26T07:28:43Z WARNING: _cleanupRequests: Asset download request timed out for ed124764-705d-d497-167a-182cd9fa2e6c.sound 2007-06-26T07:28:43Z INFO: assetCallback: Boom, error in audio file transfer: File transfer timeout (-23016) 2007-06-26T07:28:48Z INFO: Exiting main_loop 2007-06-26T07:28:48Z INFO: disconnect_viewer: Disconnecting viewer! 2007-06-26T07:28:48Z INFO: LLInventoryModel::saveToFile(/home/paul/.secondlife/cache/d22c033a-ee70-448b-8607-981374f13367.inv) 2007-06-26T07:28:48Z INFO: LLInventoryModel::saveToFile(/home/paul/.secondlife/cache/ba2a564a-f0f1-4b82-9c61-b7520bfcd09f.inv) Segmentation fault (core dumped) ok, new warning is: So I have to go to: -hum, that's not enough, still same problem, guess I will need to logoff, then relog on my account -had to reboot because of a window creation error
-The warning about pulse-rt group is gone -still no success on ESD initialization: 2007-06-26T07:48:09Z INFO: LLAudioEngine_FMOD::init() initializing FMOD 2007-06-26T07:48:09Z INFO: init: Trying ESD audio output... 2007-06-26T07:48:09Z WARNING: ESD audio output FAILED to initialize: Error initializing output device. Yeah!: no more: /bin/sh: /usr/bin/esd: not found error -Tried to add pulse and pulse_access groups to myself, but that did not help
-So I finally tried Liandra Caewlin ALSA-OSS suggestion (step 1 & 2) -watch out editing this long line, you may end up breaking it, and secondlife would just seems to hang without opening a window (guess it depends on where you break it) -you will need to 'sudo apt-get insall alsa-oss' -Now I got A LOT of 'FSOUND_Output_OSS_Wait : Timeout on audio write. Caused by bad driver!' But I guess this have to do with the driver for High-Definition Audio (Intel AC97 replacement) Result: Thanks Liandra! This issue was drastically edited from its original version, and contains a lot of confusing information in the comments, because it's not clear if the comments are referring to the description as currently written, or as previously written, unless you read it extremely carefully.
If there is still an issue to be resolved, please open a new issue in the correct project (e.g. if it's a Linux viewer only issue, then it belongs in VWR, not SVC). Please do not reopen this issue (only link to it if necessary). |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
Bug
Resolved
Showstopper
Login -> Freeze -> LLCircuit::addCircuitData
