Jump to content
Sign in to follow this  
Oskar Linden

Deploys for the week of 2012-08-13

Recommended Posts

I'm seeing some very unusual behavior with the latest LeTigre. Second Life RC LeTigre 12.08.13.263365  Siim Name: Borgbeef

 

Sim restarted fine, and I returned once it came back up. I created a couple cubes to test some of the fixes mentioned in the notes, then when done tried to delete them.  I started getting the error "Second Life: Cannot derez object due to inventory fault." over and over.  So I decided to log out and back in since I wasn't able to start a group chat.  I then discovered I could not log back into Borgbeef, the viewer would give the "Failure to connect to a simulator" message.

So, I logged into the sim next door, and discovered Borgbeef was 'invisible' but still up. Attempting to cross in caused the error " Second Life: Failed to grant capabilities".   While explaining this to the Beta group, it suddenly became visible.

So I crossed into Borgbeef, and tried deleting the two objects that failed originally. They deleted this time. However, the group chat again died after entering the sim, so went to cross back into the other sim, only to get this:

http://home.comcast.net/~volfin/20120815183249320.jpeg

I was able to TP out however.

I"m still finding it's impossible to start group chat sessions or to particpate in group chat while in that sim only.

 

Now that some time has passed, things seem normal. It may be this was related to this JIRA:  https://jira.secondlife.com/browse/SVC-8019  It's certainly the first time I've had any of these issues happen after a restart, and I've been doing this since LeTigre came to exist.

 

Share this post


Link to post
Share on other sites

Well Oskar, SVC-8124 and its relatives are now living wild and free on Main Server and LeTigre sims.  SL is no longer a safe place for anyone with a limited Download budget.

This is beyond sad. It should NEVER have been allowed to happen.

Straws and Camels' backs spring to mind.

And Yes I have posted both on SVC-8155 and SVC-8124 and included a log file.

NOW.  FIX THIS (that message is aimed at Linden Lab, not you personally, Oskar)

Share this post


Link to post
Share on other sites

Anything that uses the typical trick "llSetLinkPrimitiveParamsFast (LINK_THIS, [PRIM_FLEXIBLE, TRUE, 0, 0, 0, 0, 0, <0,0,0>]);" to make a single prim phantom in the linset stop working. This includes phantom doors and many sculpted buildings using a script to make phantom the door frame.

Share this post


Link to post
Share on other sites

I've seen what Linden Labs try to do being called Agile Development.

Ever since Pathfinding started to be tested in the Magnum RC, it's looked as though somebody had tied the developers' bootlaces together. It feels like there has been some essentially political pushing going on, and I've seen one claim that it is real bleeding edge technology, something nobody else is doing with Havok.

What seems to have happened, from what those of us outside the Labs know, is that something associated with Pathfinding triggered several similar bugs, with the same effect of vastly increasing traffic between servers and customers, enough to bring into play usage caps and extra charges after what a customer's ISP considers excessive use.

Forcing the rollout, and the promotion of RC code to Main Server, before these bugs could be dealt with, sounds about as agile as dropping a sixteen ton safe on somebody's head. It looks like full-blown Wile E. Coyote agility. There were signs of bugged code in all RCs reported before the weekend.

The First Law of Cartoon Physics seems to apply here.

Oskar is working with a system that is fundamentally broken. There isn't always time to react before the next roll-out is due, and it now seems to be that roll-outs MUST happen on schedule. Pathfinding, it seems, has been too big and complicated to be squeezed into the work patterns the Labs use for new server code, and nobody has been willing to amend those timetables.

It hasn't helped that other things have been going on. The new tools for carrying out a code roll-out still seem incredibly slow (but some of the apparent lethargy could be down to things such as NavMesh creation). I haven't seen any reports I would want to rely on, but I understand that the labs are shifting from three server centres to two. That could make it harder to spot the network traffic spikes, because there isn't the established baseline for the new topology. It also can mess up sim-crossing for a while.

OK, there was also a huge amount of network traffic between the UK and the rest of the world during the Olympics. I expected a certain general crankiness, and saw it.

Share this post


Link to post
Share on other sites

I'm afraid that is nothing new: side-effect of the Pathfinding system, and using that behavior is an "unsupported hack"

The Lindens are not very good at telling people things. I cnnot point to any official statement on this.

 

Share this post


Link to post
Share on other sites


EsePibe Borkotron wrote:

Anything that uses the typical trick "llSetLinkPrimitiveParamsFast (LINK_THIS, [PRIM_FLEXIBLE, TRUE, 0, 0, 0, 0, 0, <0,0,0>]);" to make a single prim phantom in the linset stop working. This includes phantom doors and many sculpted buildings using a script to make phantom the door frame.

Yes, that hack was removed for Pathfinding compatibility.

New method is to set the whole Linkset to Physics Type: "Convex Hull", then choose the indivudial prims in the linkset you want to be phantom and set them to Physics Type: "None"  The "Features" pane of the edit tool has this setting.

Share this post


Link to post
Share on other sites

Oskar

I find it remarkable that at this time, more than two full days after the initial posting of this thread and more than 24 hours after the RC rolls, the Release Notes for the RC channels are still not available.

Some folk might ask "What are they hiding?".  I just ask, and not for the first time; "Where are they and when will we be allowed to see them?" 

Share this post


Link to post
Share on other sites

woo in without signing in...spooky...

Well, whatever...Today I reported 5 sims in the Blake Sea and Nautilus areas as being offline, even though my Ban-line HUD and the SL map showed them online. 5! Turnbuckle, Half Hitch, Murat, Nautilus Symacom and Nautilus Blyth!

SL is falling to pieces.  Never mind the sky falling...there's going to be nothing for it to fall onto soon.:smileytongue:

Share this post


Link to post
Share on other sites

When something is introduced that stops someone being able to log on to Second Life because an update has caused them to use their montly allowance in just 3-4 days, then it needs to be addressed. Or will  all those people stopping off the grid right now until this is sorted be fully reimbursed for their membership and tier which they cannot use?


I rather suspect the answer to that is no, of course not, but unless this is resolved asap then more people will be forced to leave. It might only be 'on' Magnum sims, but it certainly affects all bar private islands. And if you pay a tier for mainland, that's where you want to be able to go.


More info on this please  - it's not really enough to say 'it's being looked at'.

Share this post


Link to post
Share on other sites

Lindens are obviously in a panic about this I contacted  live help a couple of nights ago and asked for pathfinding to be turned off in the regions where I own land...quite remarkably I was offered an island without paying the 1000 dollar set up fee...wish I could afford the monthly fee for an island ....

 

It was also suggested I use the sl viewer instead of firestorm ...didnt make any difference and to set the viewer distance to 32...odd it only seems to go down to 64 on my version... but then again if Im building with a 64 m prim I at least want to be able to see the other side of it to build with in...

 

In any event I cant get into sl while this fiasco goes on...not without paying my ISP a very large amount for going over my gb allowance. Of course even though I can only use sl for a limited amount of time Im sure oskar linden wont be recommending refunds on both tiers AND premium payments....

Share this post


Link to post
Share on other sites

Oskar Linden is clearly the kind of guy who wont admit a massive error has been made and that users are being inconvenienced BIGTIME...

 

EXTRACT A DIGIT OSKAR AND GET THIS USELESS FEATURE TURNED OFF I WONT BE WAITING ROUND FOREVER FOR THIS ONE TO BE FIXED...Not like I have with numerous ludicrous boundary crashes for the last 5 years....

Share this post


Link to post
Share on other sites

Nothing personal, and yes having SL suddenly use more bandwidth than usual is not a good thing, but it's MY responsibility to monitor MY data useage. I do monitor my data useage across several devices and I don't whine at the provider of a game or app if I let it use up my data allowance. That's MY responsibility. Geting careless and blaming anyone else is immature.

Share this post


Link to post
Share on other sites

Well done for reporting them as down Ayesha, but as you say, how can you tell they are down. It did seem there were quite a few sims down but they didn't show up as being down and there are other strange things too. For example with Blake Sea - Black. I can walk in to that sim from a sim next to it, but I can't TP in to it or get in to it with a plane. Strange !

 

Share this post


Link to post
Share on other sites

Cincia

That is a pointless statement.  The bandwidth use change is not a small one it is in some cases x100.  Many of the folk affected are not careless or immature.  They are not whining, they are complaining justifiably.  In my view.

Share this post


Link to post
Share on other sites

I have reported several sims as down in the last week or so. It's easy to do.

I usually include a short description of what I see, which is usually a water region which is showing as blank on the mini-map, and no access from an adjacent region.

Maybe it is some other fault, but if it walks like a duck and quacks like a duck, I figure it weighs the same as a duck.

 

Share this post


Link to post
Share on other sites

so when your SL usage went through the roof you did what? This is a very narrow minded view. This is a problem which affects a large number of people. I'm quite sure the majority don't check daily, but they do check  monthly - by which time they will suddenly be faced with a huge bill which they've never received before - simply because LL didn't check something again.

Blaming users for programmer mistakes is all too common. How about giving users a break too - they will check, and they most certainly should not be given a bill for £200 because somehting they've used with no problems for years and months has suddenly gone through the roof

Share this post


Link to post
Share on other sites


Cincia Singh wrote:

Nothing personal, and yes having SL suddenly use more bandwidth than usual is not a good thing, but it's MY responsibility to monitor MY data useage. I do monitor my data useage across several devices and I don't whine at the provider of a game or app if I let it use up my data allowance. That's MY responsibility. Geting careless and blaming anyone else is immature.

 

My own personal viewpoint is that I will be bloody annoyed if I have to severely limit my usage of Second Life, simply because bandwith goes up a ridiculous amount for a feature I personally could live without.

I am aware that LL can do this, can do what they like, make changes, etc., etc. I am aware also of my own personal responsibilities, but for a large number of residents, especially those who are charged not a flat-rate for their internet, it's going to be a big kick in the wallet.

Share this post


Link to post
Share on other sites

Actually, although I don't know the legal position in the USA...

Such a change by a service provider such as Linden Lab without due communication of said change could be argued as a negligent business practice.

That would get Linden Research's attention!

Share this post


Link to post
Share on other sites


Ayesha Askham wrote:

Actually, although I don't know the legal position in the USA...

Such a change by a service provider such as Linden Lab without due communication of said change could be argued as a negligent business practice.

That would get Linden Research's attention!

You'd need a good lawyer because we don't have a law for it specifically, but if your bill goes up high enough due to this (without a word from LL warning their users) LL will be liable.

The problem with that solution is good luck ever getting back into SL again afterwards.

Share this post


Link to post
Share on other sites

What I find very strange is that the relevant JIRAS on this issue (there seem to be two and a nice little argument back and forth as to their purity / dupliaction) are not particularly high in terms of votes or being watched,


If I read the various comments correctly, at least a few people have had ISP billing or cap issues already.   Given the very wdiespread nature of the problem, Id have thought more people would be affected by now.   Left unchecked SL viewership numbers would be down.


In my own case, while I am seeing at least one order of magnitude increase in bandwidth (unless I throttle DD right back to well under 100m) I am not seeing humngous increases in my daily bandwidth as reported by ISP.   Its hard to be preicse as other things are going on with my connection, but most of the time SL probably dominates (I do very little media streaming at the moment).


So - heres a question - I assume that the statistic most people are complaining about is BW as reported by the stats bar.   Is there any possibility the stat is being reported incorrectly?


There seems little doubt that spam is up.  But I wonder if that can account for the apparent BW increaes.   3 or 4 MBPS is a LOT of spam (text).

Another fact is that I have stated getting a lot of client crashes (Phoenix).   Perhpas that is due to the spam ... my log file has hit 3MB several times.

Also we have several differnt flavours of spam, it seems.  Common cause, or not?

Just wondering if this problem(s?) is more complex than it (they) seem(s).

I am most intgerested in how many people are seeing actual increases in their BW as measured by the ISP.

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Sign in to follow this  

×
×
  • Create New...