Has support ended? Does Second Life slowly fade?

[Moni Duettmann]

I'm writing this, because I wonder what others think. I have always been a "good" Secondlifer in the regard that I have been joinng technical discussions on the JIRA website, reported bugs, tried to provide to solutions for years. Yet I think that fewer and fewer questions are answered, bug reports get "closed" although no solutions are yet provided. In this case it's about an old old issue, the rezzing of textures, which stiil in all my current viewers doesn't work properly. Some texurues don't load at all until I click on them, others keep switching between steps of uncomplete loading, some get blurry again after having fully rezzed before. On the jIRA website I see many reports about this, but no changes happening. Most reports have simply been closed without solution. This happened to my own report as well. When looking at the few answers that went in before the closure, it makes me angry. Users and Linden staff are seriously trying to help with suggestions like "did you try to clear your cache"... oh my... Before I even take the effort to file a bug report, I have been going through all the standard procedures designed to bring possible soultions. BTW: those never do. At least I can't remember a situation where "clearing the cache" solved a problem. Still this is all the support that is offered. To me this is so frustrating. Second Life without textures is just not possible. This should have been solved years ago. So where is real support? I don't see it! I still pay my 72.-$ every year (which is more than to any other internet service I use), but I hardly use Second Life anymore, because this way it just doesn't make sense. Wonder what other people think about this!

[Stefanosje Sellers]

Well, Moi, this thread is only going to ask for a hell of a lot arguments against/defending LL. I've never thought of SL support as brilliant, however I can assure you most of the support-desk employee's of companies like LL give you such responses. They go by the book and only have a small set of solutions they can offer, after which they decide if they will pass your ticket on to line 2/3 (terms for mid-tech and programming department).

I think it's safe to say LL doesn't care anymore, however.

[Ren Toxx]

The textures issue is actively being investigated, with two distinct improvements planned - and thus probably to be implemented sooner or later. Read Inara Pey's blog coverage of Peter & Oz Linden's live interview on "The future of Second Life".

[Qie Niangao]

---

Moni Duettmann wrote:

At least I can't remember a situation where "clearing the cache" solved a problem. 

---

This is so weird. I keep hearing folks say this, but I get corrupted texture cache about once a week, and it fixes by clearing cache -- always and only by clearing cache.

I am almost always using one or another beta viewer, so that may contribute to more cache corruption, but still it's so odd that some folks seem never to have the problem.

[Orca Flotta]

---

Qie Niangao wrote:

---

 I get corrupted texture cache about once a week, and it fixes by clearing cache -- always

and only

by clearing cache.

---

Yes Qie, corrupted taxtures is the only known problem that can be fixed by clearing cache. Nothing else actually.

But OP's problem sounded different, like connectivity probs, wrong graphic settings or wrong bandwith settings. Didn't sound like  corrupted textures at all.

[Perrie Juran]

---

Qie Niangao wrote:

---

Moni Duettmann wrote:

At least I can't remember a situation where "clearing the cache" solved a problem. 

---

This is so weird. I keep hearing folks say this, but I get corrupted texture cache about once a week, and it fixes by clearing cache -- always

and only

by clearing cache.

I am almost always using one or another beta viewer, so that may contribute to more cache corruption, but still it's so odd that some folks seem never to have the problem.

---

I'm surprised that you don't install them in different locations to avoid cross contamination.

But something strange did happen to me last week.  I crash a lot because of the texure memory problem.  Normally I just relog.  But last week FS could not get pass "loading texture cache."  I wound up having to do a reinstall.

[Ohjiro Watanabe]

Make sure you are using HTTP for receiving textures. In Firestorm the option is in Graphics/Rendering. HTTP has error checking which will help prevent corrupt cache. If this option is not checked it means that you are using UDP for texture download which doesn't have error checking and may well corrupt your cache.

[Qie Niangao]

---

Perrie Juran wrote:

---

Qie Niangao wrote:

---

Moni Duettmann wrote:

At least I can't remember a situation where "clearing the cache" solved a problem. 

---

This is so weird. I keep hearing folks say this, but I get corrupted texture cache about once a week, and it fixes by clearing cache -- always

and only

by clearing cache.

I am almost always using one or another beta viewer, so that may contribute to more cache corruption, but still it's so odd that some folks seem never to have the problem.

---

I'm surprised that you don't install them in different locations to avoid cross contamination.

But something strange did happen to me last week.  I crash a lot because of the texure memory problem.  Normally I just relog.  But last week FS could not get pass "loading texture cache."  I wound up having to do a reinstall.

---

There hasn't been much chance for cross-contaminating, but it's a good suggestion and something I should be careful about going forward. Until very recently, there wasn't anything left to contaminate because I kept cache on ramdisk, but I've just changed operating systems and  cache now persists on the SSD, so it could conceivably happen the next time the viewer upgrades

Incidentally, about that texture memory problem: One of the reasons I finally switched to Windows after over a decade of running Linux on the desktop, was an attempt to escape that same texture memory problem -- at least, that's how it was reported in the SL console log. It took me about a month of trying everything (including 64-bit Ubuntu -- which I do not recommend) before I discovered that I had a bad stick of DRAM. This despite the fact it always and repeatedly passed every stage of memtest86 with flying colours.  Both Linux and Windows 7 worked without a hitch, but what finally gave it away was that Windows 8.1 simply refused to boot with that memory present but worked fine when it was replaced with identical memory borrowed from another machine. 

So, apparently those texture cache problems might be indicating a hardware problem too subtle to appear with normal system diagnostics -- nor even targeted tools such as memtest86. At least I haven't seen them recur since I got rid of that almost undetectably bad memory.

(Note: It is absolutely true that the bad memory could have caused cache corruption, but I've had it occur a couple time already since I fixed that problem. I should explain that I'm not particularly complaining about cache corruption as a problem, but more puzzling over the mystical attachment some folks feel for their ancient cache contents.)

Oh, and yeah, I've used HTTP Textures constantly for a long time. When they were first an option, I'd toggle that off and on to get stuff to load, but that was a long time ago, and now loading performance with HTTP Textures is just head-and-shoulders better than without it.

[Perrie Juran]

---

Qie Niangao wrote:

---

Perrie Juran wrote:

---

Qie Niangao wrote:

---

Moni Duettmann wrote:

At least I can't remember a situation where "clearing the cache" solved a problem. 

---

This is so weird. I keep hearing folks say this, but I get corrupted texture cache about once a week, and it fixes by clearing cache -- always

and only

by clearing cache.

I am almost always using one or another beta viewer, so that may contribute to more cache corruption, but still it's so odd that some folks seem never to have the problem.

---

I'm surprised that you don't install them in different locations to avoid cross contamination.

But something strange did happen to me last week.  I crash a lot because of the texure memory problem.  Normally I just relog.  But last week FS could not get pass "loading texture cache."  I wound up having to do a reinstall.

---

There hasn't been much chance for cross-contaminating, but it's a good suggestion and something I should be careful about going forward. Until very recently, there wasn't anything left to contaminate because I kept cache on ramdisk, but I've just changed operating systems and  cache now persists on the SSD, so it could conceivably happen the next time the viewer upgrades

Incidentally, about that texture memory problem: One of the reasons I finally switched to Windows after over a decade of running Linux on the desktop, was an attempt to escape that same texture memory problem -- at least, that's how it was reported in the SL console log. It took me about a month of trying everything (including 64-bit Ubuntu -- which I do

not

recommend) before I discovered that I had a bad stick of DRAM. This despite the fact it always and repeatedly passed every stage of memtest86 with flying colours.  Both Linux and Windows 7 worked without a hitch, but what finally gave it away was that Windows 8.1 simply refused to boot with that memory present but worked fine when it was replaced with identical memory borrowed from another machine. 

So, apparently those texture cache problems

might

be indicating a hardware problem too subtle to appear with normal system diagnostics -- nor even targeted tools such as memtest86. At least I haven't seen them recur since I got rid of that almost undetectably bad memory.

(Note: It is absolutely true that the bad memory could have caused cache corruption, but I've had it occur a couple time already since I fixed that problem. I should explain that I'm not particularly complaining about cache corruption as a problem, but more puzzling over the mystical attachment some folks feel for their ancient cache contents.)

Oh, and yeah, I've used HTTP Textures constantly for a long time. When they were first an option, I'd toggle that off and on to get stuff to load, but that was a long time ago, and now loading performance with HTTP Textures is just head-and-shoulders better than without it.

---

That's absolutely fascinating, especially about the Ram Sticks.  If I had some laying around I'd sure give that a try.

My Computer and Second Life have a convoluted history.  It is a custom built gift from family.  It was humming along as happy as could be for about a year until Mesh went live.  The first official release Mesh Viewer brought it to it's knees......5 FPS in a crowd whereas I was used to getting 20.

I was running XP and one of the odd things is it always reported "CPU: AMD Processor model unknown (3000.13 MHz)."  The Linden who worked on my (now hidden, wtf) JIRA had me try a ton of different things.  I was also not the only one having the problem.  As it turned out it was a change in the way flexi's were handled causing it.  There were some updates that helped some people but not me.  The Linden working on it said he saw "unknown processors" all the time and that it was not a concern.

Another funny thing was I consistently got better performance with Firestorm, but no where what I was used to.

After sufferinf with this for about a year a friend made me a deal on an extra Win7 licence they had.  And poof, all my problems went away.  I did still have the "unknown processor" and dug into it one more time.  I got out the install CD that came with the processor and it had the same drivers as what Windoze had installed, a very old generic Windows driver.  I dug through AMD's boards and as best as I could determine, AMD never wrote a specific driver for my processor (AMD Athlon II X2 250). 

Everything was fine for a few months but then after a new viewer release (I forget which one) the crashes started.  With out dragging this out further, this was the Lab's response:  "But we do not think it is a bug."

All I can say is how can you have a memory problem in the Viewer that afflicts so many people and say, "We do not think it is a bug?"

 

[Moni Duettmann]

Ren: I'd be super-happy, if that's the case. I remember this problem happening even in early times, like 2007. Then it disappeared as if being solved, but it returned more than a year ago.

@Qie a.o.: I don't think this is a problem of corrupted textures. I *once* had a corrupted texture, that had to be replaced, so I know what a corrupted texture is. This is different. Textures do load, but then they unload and get blurred again. 

As for the hardware suggestion: In another forum someone suggested that it could be related to the graphics card, resp. driver update. Since I am using one of the most ordinary cards, which is built into many Mac books, Mac Minis et al, and using the actual system update (which includes graphics card driver update on a Mac), I don't think it's that. Actually with the same Mac I didn't experience these problems two years ago. The bug came back after this little wave of viewer updates related to the introduction of this new avatar rendering thing (forgot how that was called), when all Viewers, incl. Phoenix, Firestorm et al had to relase new versions.

One of the recommendations you regulary get, when reporting a bug is, that you should extend your RAM. It's one of those things that at least for me never made any difference, neither in resolving bug probhlems, nor in general increase of framerate or else. It's just one of these things people say.

I often hear: lower your draw distance. Hm. Of course I can do that, but it doesn't help with the texture bug. I mean, when I am in a room, where all walls are not further than 10m away, the textures already don't load, no matter what my draw distance theoretically enables behind the walls. The only thing that you can trigger with toggeling draw distance, is your frame rate.