Ayesha Askham

Hi again Oskar So, now that the new Release notes are up....Still I see nothing new. Are we getting revised software? Maybe I should know but I don't see it from the notes! Looks like a "wait and see" job to me!:smileywink:

Hi Oskar Just a query. I just looked at the release notes for LeTigre, and the notes only refer to the last server version, not that which is slated for tomorrow. Could you check them please? Or did I misread and there is no change? Ta, Ay

Y'know I'd buy into Rodvik's hubris if it wasn't so misplaced. Sim crossing is nowhere near as good as it was Pre-Windlight, I think 10% would be generous. Sailing on Blake Sea has become a challenge.."How many crossings can I make before my boat derezzes/I crash". This, considering how well-publicised SL Sailing is isn't merely SL footshooting...it's commercial suicide. Over the past 2 years Linden Labs have ostracised Adult activities and BDSM, gamers who lost HUDS in summer 2010, and now SL sailors as well as racers. It's barmy. I'm beginning to wonder what's left? As to spending more time devising new shinies...it is blazingly obvious that Linden Lab developers and programmers would rather devise new toys than fix broken ones...where's the fun in toy-mending? Until the company actually start to understand that SL actually has to WORK to succeed, this slow atrophy will continue. It's not rocket science, it is simply good business!

@Oskar As promised, an update: About 10am PST our sim was rolled as part of the LeTigre server update. The situation is now: You are at 257370.0, 321842.0, 901.8 in Woods of Heaven located at sim21110.agni.lindenlab.com (216.82.30.144:13028) Second Life RC LeTigre 11.11.04.244558. So we are still on LeTigre. However the strangeness does not end there. I have a "Spray" sailboat, made by Lia Woodget, version 5.5, which, after the reintroduction of physics functions after Andrew's security patch killed the existing physics function, sped around our sim like a speedboat, much to the confusion of its creator and the scripter. Nothing they could do seemed to correct the matter and I could reproduce the effect a couple of weeks later on Magnum sims (at Blackspot Shipyard). Now it is behaving perfectly again. So I ask....why? Is there something in the new release on LeTigre that would explain it? And moreover, is the boat going to fail again when (and if) we return to the Main Server? For now, notwithstanding a revelation as to "what gives?" I would like our sim to stay on LeTigre, since I have seen none of those nasty lag spikes in the last 24 hours. Is that OK with you, Oskar?:smileysurprised:

@Oskar Thankyou, Oskar for that open and rapid reply. I know you and your team must be dreading Restart day now, but that explanation is more than good enough! This issue of server-load is a very germane one to a lot of SL's current woes, perhaps your masters could see that you really cannot squeeze a quart into a pint pot without a lot of spillage and mess. I'll keep you posted about our server host and channel after the LeTigre restarts today.:smileyhappy:

Oskar Now, everyone knows that things are not going well with the roll again...but.... While I realise that we are lucky to be back online at all, what is the rationale behind moving our sim to LeTigre? "You are at 257370.0, 321842.3, 901.8 in Woods of Heaven located at sim21110.agni.lindenlab.com (216.82.30.144:13025) Second Life RC LeTigre 11.10.28.244134." We were on sim 5338 Main Server and had been on 4 or 5 thousand series sims all year...now 21110? Please don't anyone think I'm bitching, I'm just pleased to be online, 'cos running RLV if my home isn't online (where I last logged out) I cannot log in! I just wonder why LeTigre?

Just putting this in to see what response it gets. Recent performance issues on Full-region sims appears to be covered by JIRA #SVC-7339 but Maestro Linden was at pains to point out to a couple of posters there that that JIRA was NOT concerned with the "Lag-Spike" issue on Homesteads. Does anyone know if there is a JIRA specifically related to that phenomenon? If so I'd like to post to it,'cos this issue is getting on my nerves. If there isn't one (and there ought to be), I'll start one. Edit: Seems to be SVC-7423. Also, our sim was restarted to LeTigre from Main Server today...we are one of the lucky ones to actually get back online after yet another painful restart experience. I wonder if Oskar is bald yet?

Just putting this in to see what response it gets. Recent performance issues on Full-region sims appears to be covered by JIRA #SVC-7339 but Maestro Linden was at pains to point out to a couple of posters there that that JIRA was NOT concerned with the "Lag-Spike" issue on Homesteads. Does anyone know if there is a JIRA specifically related to that phenomenon? If so I'd like to post to it,'cos this issue is getting on my nerves. If there isn't one (and there ought to be), I'll start one. Edit: Seems to be SVC-7423. Also, our sim was restarted to LeTigre from Main Server today...we are one of the lucky ones to actually get back online after yet another painful restart experience. I wonder if Oskar is bald yet?

Oskar Approaching a holiday?? I know you folks in the USA have Thanksgiving this month, but Christmas (or whatever cobbled-up non-religious name the PC mob wish to give the time) is six weeks away! I suppose there is Hanukkah, Diwali is past, as is Ede, but come on guys, how many weeks' holiday do you get?:matte-motes-big-grin-wink: No, I just want this TD spike issue sorted out, and as soon as possible, and Yes...I will be patient while you do it, just ...please do it! Edit: OK so I can't count!:matte-motes-zipped:

Oskar Approaching a holiday?? I know you folks in the USA have Thanksgiving this month, but Christmas (or whatever cobbled-up non-religious name the PC mob wish to give the time) is six weeks away! I suppose there is Hanukkah, Diwali is past, as is Ede, but come on guys, how many weeks' holiday do you get?:matte-motes-big-grin-wink: No, I just want this TD spike issue sorted out, and as soon as possible, and Yes...I will be patient while you do it, just ...please do it! Edit: OK so I can't count!:matte-motes-zipped:

I put this up as much in the hope of it being shot down as in the hope of getting some sensible answers. To a degree I'm playing Devil's Advocate and proposing something unlikely (I hope). We all know that SL has been subject to a worsening TD (lag) situation for a few weeks now. It comes and goes but seems to worsen with increased concurrency (more folk logged in). We also used to know that several regions shares a server...is it 4 full regions to a server core or 4 full regions to a quad core server? Either way a Homestead gets one quarter of that allocation, and Open Sims half again. Now, those with more than a year of SL behind them will recall the days when the sim stopped for several seconds when an av logged on to a sim or TP'd in. That was overcome, largely by limiting script-tolerance and improved software. We recently had a server Operating Sytem upgrade. Is it possible that there has been an increase in crosstalk between sims on a server or on a core? Is is possible that when an avatar logs into a sim somewhere or TP's in, other sims on that server or server core lag? Would that also go some way to explain the atrocious sim border crossings we suffer presently? If that were to be the case, what would be the cure? Better software? Debugged OS? Or lower script limits? In my more Macchiavellian moments I wonder if, rather than announce lower limits publically, with all the negative connotations that would have, whether Linden Lab has simply allowed a change in the basic structure of SL to bring about the result they wish to see? Is that too cynical to be true...I hope so.

Ishtara Many sims are suffering the same symptoms as you report, and so far I have not seen a coherent comment from either a Linden or one of their trusted confidants. The issue is not confined to Full-region sims, though I imagine it is more noticeable there simply beacuse such behaviour is not expected there. Homesteads are most certainly seeing reduced script-tolerance, and have been doing some ever since those borked security fixes which killed SL physics functions. Those functions were eventually restored, but something changed in the way scripts and physics objects interacted with the servers. With new quad and multi-core servers, the density of sims per server must have increased in the past couple of years, but that has not been shown in a reduction in tier paid by Full-region or Homestead sim owners. While it is true that the memory and processing speed of a single core is equivalent to what was a single processor a few years ago, they are not identical, and especially in the case of interlinked entities such as SL regions, the potential for crosstalk does not seem to have been appreciated. As to whether Linden Lab have sought to squeeze even more sims onto fewer servers - that would be very dangerous territory for Linden Lab to enter, since there could be a case made for misrepresentation, in European law, if not in Federal law. The use of a restart to clear a clogged memory is, as you so rightly say, a very risky business in the current state of SL, and yet many Homestead owners are exhorted to use this course as an anti-lag tool, by individuals that really ought to know better. Script-limits, on either Homestead sims or avatars would certainly reduce the impact of this shortcoming, but it would not tackle the fundamental failure of the system that seems to be occurring. To say that imposing stringent script-limits on Homesteads would be extremely devisive and short-sighted of LL would be both an understatement and so typical of Linden Lab. PS: On reflection, perhaps these issues that we are seeing are, in fact, a covert operation by Linden Lab to cause a reduction in script-use by residents. Perhaps this increased script intolerance is intentional? If that is the case, I would posit that a publically announced intention to reduce script allocation would have dissuaded the early-stage residents from remaining in SL, since they will have come to appreciate what script-functions can do, but will not have acquired the skills to descript or modify such articles as they might have. I'm sure that such a thought is too paranoid and conspiracy-theoretical to carry much weight.:matte-motes-big-grin-evil:

@Oskar, or anyone that might know from the community: 1) There is no mention of the RC Channel roll finishing on the Grid Status Page - has it finished is is work still ongoing? Also 2), something that I've observed in the last 2 or 3 weeks: After a restart our sim (a Homestead) has seemed OK, running at TD better than 0.97 and ping at about 140ms, but after about 24 hours we start seeing spikes of TD as low as 0.3, more often 0.65, with sim pings of several seconds. These phenomena are very short-period transients, one or other of our scripted animals goes into low earth orbit, and I see some chatlag. We did have a major server crash a few weeks back that showed some similarities, and I was told unofficialy that it was TIMEWARP. We were moved to a new server, and all seemed well. Now we are moved every scheduled restart (this NEVER used to happen) and these little spikes seem to be a constant companion now. I just wonder if anyone else sees this effect, and if anyone can shed any light on what might be causing it? I should point out that our sim is carefully audited to keep scripts to a sensible level, and others on the sim are experiencing this issue, it is not tied into a specific ISP, router or PC.

@Cincia The issue here, Cincia, is that a restart every couple of days is not wise for two reasons: 1) It increases the risk of a faulty shutdown which can result in data (ie object) loss. 2) It is only a band-aid, not a cure. If a sim is needing a restart every few days, and all other aspects of that sim have been carefully audited (ie scripts using excessive server time etc), it is diagnostic of a systemic issue with SL, either in the Server software or (please no) the underlying Operating System. Yes, of course sim managers or owners should be diligent in their use of all the land tools that LL give them..and that is quite an arsenal now..but you cannot make a silk purse out of a sow's ear, no matter how many times you restart a sim.

@Wolf That, Wolf, is sufficiently enigmatic to be downright Gnomish! I think we owe Oskar and his team a vote of thanks, as was said above, for the hard work put in to get the recalcitrant sims back up (ours was one). If (and it is a BIG IF) the dying sim issue is quashed, we will be immensely grateful. You are right of course, Wolf. More testing should be done and longer periods between leaps in features should occur, but the Lab's staffing and policies being what they are, it won't happen. We both know that!:matte-motes-wink:

@Oskar Woods of Heaven went down for the restart 30 mins ago,and is not back up yet. Just adding the name to the list so's you know. I guess the team is a bit frazzled by now!

I see Cincia advocates a region restart every couple of days......Hmm. Not all that long ago I can remember hearing that a region restart every week kept the server doctor away... I'm thinking I see a trend here - does the term "Memory Leak" chime with anyone? I think someone needs to go through the code with a magnifying glass and figure out where this issue (memory leak or not) is.

@Talia We too have seen this issue on our Homestead, twice. On one occasion the server simply ground to a halt every time a script was called (one that in normal circumstances has barely any impact on server performance). On the third sim crash I called Live Chat, and was told that "A team were investigating the issue at that moment". We were moved to another server and for a couple of weeks we had no further problems. Last week the issue returned, and although the sim didn't crash, I was able to observe sim pings of over 1 second on about half a dozen occasions. The following day (Tuesday) was restart day and I noticed on the restart we were moved again (this does not seem to happen routinely on Server update restarts). I was given to understand that this issue was TIMEWARP, and that it would be addressed by the OS upgrade. That does not seem to have been the case, since server perfomance is in many cases just as bad. Expecting Linden Lab to withdraw Mesh and its supporting software is a vain exercise...it will not happen, not least because Philip Rosedale promised it, and that makes in a ""shiny" shiny"". Too much has been invested for that to even be an option. Linden Lab need to sort this one out, because from being a showstopper for some, it is rapidly becoming a Gamebreaker. PS: I should add that I don't believe that Linden Lab are doing nothing about this issue. I think they are doing what they often do, and that is running in "Full Stealth" mode, because they do not wish to appear not to be in control of their creation - bad for business confidence. In one narrow way, that is understandable, but it results in many frustrated and unhappy users of SecondLife. Sadly for us, Linden Lab does not place our discomfort very highly on its values list.

@Void I take your point, Void, and have checked the proximal end as far as I can. The PC in question is as well configured for SL as I can sensibly manage, I daresay I get better average performance than most, I'm just curious.

That was my thought Void, I was just wondering if anyone else was seeing the effect.

I should be a bit more transparent in my question. I know the phoenix recommendation, and I tested my dowload speed to the Dallas Fort Worth and San Francisco servers, to get a "feel" for the true line speed from SL to my PC in the UK. If the results are to be believed there is a huge difference between Dallas and SF, but both gave me sufficient speed to run 1500kbps as my bandwidth. It just seems to me that the transmission speed from SL is decreasing as time passes, and I wonder whether in fact SL's own servers are slowing?

I'm curious to know what various people use to determine their optimum bandwidth for SL on the current server versions.

Oskar I'm curious. I have noticed over the past week that TPs back into Main Server regions from regions on RC channels LeTigre and Magnum (LeTigre in particular) result in slow texture reloading (longer "rez times") than in TPs between regions running Main Server. Is there any conceivable reason for this? I carry a relatively high script-load relative to some, but this effect is script-load independent. (Or certainly seems to be). Anyone else seeing this? PS I use HTTP Texture get - I've tried going back to UDP but it's too damn slow to load new textures.

@Nalates Firstly, it is true that your Blog is a good place to find out sometimes obscurely reported issues in SL. This begs the question. It should not be down to individual independent residents (Advisor in the Forum status notwithstanding) to disseminate information about SecondLife. Linden Labs should be doing that in a transparent and timely manner themselves. So you knew about the OS upgrade weeks before it started? It is strange then that more scuttlebutt didn't surface regarding it then. The SL rumour-mill is usually more effective! I am interested in the workings of SL, but, like so many other folk, my RL schedule simply does not allow me the luxury of attending inworld groupd such as that. I don't see why I should therefore be dependent on 3rd parties to pass on information as they see fit. Such a process could be highly selective. I know information from LL could be selective, but it has the distinct property of being Official.

@ Nalates Only a liitle whinge: it took 10 hours of that issue before the comment arrived on the grid status page. Linden Lab knew there was an issue long before the posting appeared. However, to be fair to Linden Lab, I guess they needed to be sure there actually was a problem with their system before they posted.:smileyindifferent: