Jump to content

Deploys for the Week of 2011-03-07


Oskar Linden
 Share

You are about to reply to a thread that has been inactive for 4800 days.

Please take a moment to consider if this thread is worth bumping.

Recommended Posts

Second Life Server (main channel)

This is getting the "server-maint" project from the Magnum Channel. This branch contains various bug fixes including {{jira|SVC-6781}}: "Duplicating old attachments does not copy attachment point/position/rotation info". Some people may notice a slight improvement in regions with heavy object creation/deletion.
Second Life RC BlueSteel
This branch will stay the same and get a merge from trunk with a possible bug fix.
Second Life RC LeTigre
This will get Andrew's "maint-server" branch. There was a bug during the deploy last week that is fixed. This branch contains several bug fixes, some security fixes, some fixed crash modes, and various performance enhancements. 
Second Life RC Magnum
This channel will get the "Estate Transfer Limits" project. This project implements adjusted behaviour limiting L$, IM, and object transfer for accounts who are limited to estates.

 

Link to comment
Share on other sites

Thanks for the information Oskar.  For what its is worth there have been some serious improvements over the last few weeks with most centering on Logins.  Avatar loading is not much easier, and generally can be done in one or two attempts and inventories are generally loading much better too, though there is still room for improvement.  This though is a major advance and if it holds will do a world of good.  The down side is that crash and lag off ports has become a major problem with over half the ports sent by someone else resulting in one or both people crashing.  Again since this does not happen immeiately, I am assuming it is a memory leak  and should be easy to fix.  This problem by the way is getting progressively worse and is a major problem for clubs and businesses that depend on customer mobility.  On a related note sim crossing are again slow and causing problems so I suspect you need to revisit the protocol behind both problems since they seem to be closely related as a result of the work of prior restarts.  Overall though there is improvement and opening this blog is a welcome step toward renewing communications.

 

Link to comment
Share on other sites

Just to clarify........"The down side is that crash and lag off ports has become a major problem with over half the ports sent by someone else resulting in one or both people crashing."

When you say "ports"   you are refering to  teleports??   

Link to comment
Share on other sites

Alexi, I have to say that my own observations on teleport reliability are considerably different to what you report.  Teleports and region crossings have been quite reliable and generally fast for me for some time now (with usually around 100-150 scripts in attachments).  It's possible that there has been a very small increase in disconnects for me during crossings and TPs in the last approx 2 months, compared to the period immediately before then, but the instance of that happening has been so small that it could easily just be "statistical noise".

If there's a memory leak causing you issues with failed TPs, it's almost certainly your viewer, not the server.  It's probably fair to say that pretty much all of the viewer leak memory, but the more recent ones are greatly improved over older ones.  It could also be that you have corrupted data in your local texture or object cache (the cache is stressed by TPing), so would be very worthwhile clearing all local viewer caches and seeing if that helps.  You also don't mention just what viewer you are using, it might well be worth trying a different viewer to see if that changes the failure rate for you.

As far as server memory leaks go, the indications from performance degredation over uptime on heavily loaded sims are that the leaks have been greatly reduced over the last 6 months or so.  Statistical analysis of dilation and frame rate on a heavily loaded Class 5 suggests to me that the situation is greatly improved over what we had mid-2010.

You should also consider whether your teleport failures tend to be related to entering/leaving specific sims, as an overloaded sim could easily make the failure rate far worse, particularly if it suffers badly from SVC-3895 (Mono rez lag, hopefully going to be finally addressed in the very near future).  When a sim is badly impacted by the Mono rez lag, everything essentially stops for rather too long, and that can cause TP failures due to the sim being unable to process your TP in/out in a reasonable timescale.

Link to comment
Share on other sites

Hi Oskar l own a number of mainland sims. My home sim uses SecondLife Server and our 4 adjoining sims use Bluesteel. Is there a Wiki or KB reference that would help me understand the difference between all the server types pls. l'm assuming Tigre, Magnum, Bluesteel and SL Server support different sim types? Also can you point me to a feature list for the servers and planned future changes/additional features that uses non-technical language please. Cheers  

Link to comment
Share on other sites

These are the wiki pages that we have:

 - http://wiki.secondlife.com/wiki/Beta/Magnum

 - http://wiki.secondlife.com/wiki/Beta/LeTigre

 - http://wiki.secondlife.com/wiki/Beta/BlueSteel

 

Things on the server side of things will by their very nature be more technical. I do try to describe things in ways that are easier to understand. Those pages link to the release notes for each project that we have on the live grid.

 

__Oskar

Link to comment
Share on other sites

@Oskar

Firstly, thankyou for keeping us informed, it is appreciated.  I DO tend to be critical, but I also give credit where it's due.

There are some oddities associated with this server version, but I believe there is a slight drop in lag (improvement in TD factor) with this roll on the main server software.  Some old scripts are now running smoothly again.  Perhaps they are related and the sim-server is able to service script calls more effectively now.

It is also good to see LL being pro-active in stamping on IP infringement without necessarily banning ingenuous residents. (Sorry, that's OT here).

@ Astarte:

I saw a similar issue last week when Woods of Heaven (my sim) was unreachable for 5 hours.  It wasn't offline, it just seemed that the system couldn't resolve its coordinates.  God knows why.  We were never told, it just was accessible again.  It may or may not be due to the rolls.

Link to comment
Share on other sites

Let me just add that I had trouble logging in all weekend, and when I did get in, the grid was extremely slow.  Put down a prim, and it takes 45 sec to rez.  Move the prim, and it moves back after some amount of time.  Group messages were impossible.

 

I don't know what was done on Friday.  But things have been very broke for me and my friends.  I polled them (slowly) and they were all experiencing the same thing on Sunday.

Link to comment
Share on other sites

@Shockwave, and by default Oskar

 

I know that there was some Inventory Server Maintenance again on Friday...now in my mind that has become a euphemism for "we're working on the Asset Servers....again" in Lindenspeak.  There is no doubt that the Asset servers are still painfully slow to respond to either client or direct sim server requests. Why? God knows, and he ain't telling.

I know that LL have at least started to do some IP policing and I applaud it...better late than never.  Whether this is what has fouled up the Asset Server response or what or whether it is the server-software I don't know , but some explantion is required for this...is there a JIRA yet?

Link to comment
Share on other sites

 


Oskar Linden wrote:

The Magnum code that we promoted this week has been around for a few weeks on AGNI. The changes were minimal and we don't foresee any issues. That being said please bring any critical issues to my attention.

 

__Oskar

 

Oskar,

I have become aware of an issue with RC Magnum 11.03.08.223072. I am no longer receiving Emails from IMs sent by scripted devices sent while I am offline.

It is not a spam filtering issue, nor a preferences/communication settings issue (I receive offline IMs via email from other people just fine), nor is it a failure of the code (it occurs with devices I have scripted and also standard devices such as the Magic Boxes used for Marketplace).

Please refer to the topic "My Question Was Closed But Not Solved ..." for further details.

Link to comment
Share on other sites

Darrius,

I think you are right, there is an issue with Object IM's send from Magnum regions. I have a device in Dolores (a Magnum sim) that reports some type of problems by object IM's. The last IM i got from that device was at march 9, 8:28 AM (SL time). I am quite sure it must have send a few IM's since then, but i do not find them in my email.

The same type of device in Pak (a Bluesteel region) worked without any problem and did send me some IM's in email while i wass offline.

The problem might even be worse, I also did not receive any IM from my device in the Magnum region while i was online. I am quite sure it must have send some while i was online, but i can not find them in my chatlogs.

I sure will do some testing after todays roll to see if the problem still exists.

 

Greetings, Chriss Rosca.

Link to comment
Share on other sites

You are about to reply to a thread that has been inactive for 4800 days.

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
 Share

×
×
  • Create New...