Firestorm repeatedly crashes

[MelodicRain]

Even though I'm not using the official viewer it's highly likely it'll happen there as well. I'm using latest Firestorm viewer at the moment, and it'll crash VERY frequently without any error messages. Over a period of several weeks I've observed some patterns:

• It usually crashes if I pan camera around my avatar or zoom in close to my face (e.g. when trying demos or editing my shape)
• How crowded a place is (or complexity of environment) has no effect on frequency of crashing - I can be at home alone, in a completely empty sky platform, and still crash every 30sec if I pan camera
• Graphics settings also have no effect on crash frequency - from lowest to highest
• It does NOT crash if I leave viewer minimized in background and use computer for other things, even after several hours

Another interesting thing is if I crash, it'll ALWAYS take two attempts to log back in - first attempt I would log in as a perpetual glowing cylinder, and my inventory would fail to load - second attempt would rez normally.

I was going to assume this is a memory leak issue or something but it's very odd how I can still crash in an empty sim, and how graphics settings have no impact. Granted this is an old computer, but I'm able to play at ~40-50fps with no issues on medium settings, which is acceptable.

Anyone know how I can troubleshoot this?

Edited February 23, 2022 by MelodicRain

[Qie Niangao]

I'm no expert but to get started on something, to me, this suggests a hardware problem, possibly defective memory. So the third thing I'd try would be to run something like memtest to make sure all the RAM is functioning as expected.

That's a time-consuming process, though, so also in line with hardware the second thing I'd try is something like CPU-Z HWMonitor to see if heat is a problem (which seems unlikely, based on the "40-50fps with no issues" thing, but still something is wrong somewhere).

But the very first thing I'd try is some other viewer—any other viewer—to see if there's something funky buried in one of all those myriad Firestorm settings; that's not a hardware thing but this business with "a perpetual glowing cylinder" on the first post-crash login attempt seems very odd and I'd want to repro it on some other viewer before working too hard on hardware or really anything else.

(Also, I'm only assuming this is a PC, not a Mac. Might be useful to paste in the system configuration info from Help / About. Might reveal an out-of-date driver or something.)

[tomm55]

45 minutes ago, MelodicRain said:

I'm using latest Firestorm viewer at the moment, and it'll crash VERY frequently without any error messages.

You might also find some info here: https://wiki.firestormviewer.org/fs_run_crash

There should be also a support group https://www.firestormviewer.org/support/

Edited February 23, 2022 by tomm55

[Paulsian]

Few months ago my official viewer was crashing like few times a week and then a few times a day, finally figured out that the nivida driver I updated from their site was causing my entire computer to crash not just the viewer. Did you recently have a video card driver update and it was nivida? just throwing that out there. Other than that nope never had random perpetual crashes with just the viewer before. 

[Lilac Melodious]

quite old, yes. but for Firestorm Beta 6.6.9, this issue occurs still, or randomly it will occupy 10GB of RAM. (i have 32GB) it does seem like a memory leak, which has yet to be tackled.

 

[Caly]

I'm having serious issues also - for a couple weeks now - I have gone through 3 computers and all stop when it gets to "connecting to region") poof like magic and goes no further.  After trying a few  of my older computers, they are not allowed to even download Firestorm any longer because it now says - FS is no longer compatible with my "version" of computer - maybe because they are Windows 10.  Is there someone to contact at Firestorm or is this all we can do? 

[Alwin Alcott]

25 minutes ago, Caly said:

 Is there someone to contact at Firestorm 

firestorm has it's own support on their website.

[Nalates Urriah]

A quick system file check will fix a large number of odd problems. Open a CMD window as Admin and run: SFC /scannow.

[Brises Ravenheart]

Brand new Acer ROG laptop, I7  processor with Nvidea RTX 4060, runs fantastic with sometimes as high as 160 fps. new install of firestorm, and crashing many times per day. The only clue I have is text stops posting or commands stop responding, then in a few more seconds... disconnect. This was happening with my previous laptop as well, but I had managed to get it more under control, still happened though. Not sure if this is a firestorm issue of a Linden issue. Tried the scan from the previous post, and  "Windows resource protection did not find any integrity violations."

Edited December 31, 2023 by Brises Ravenheart
additional information

[Qie Niangao]

2 hours ago, Brises Ravenheart said:

text stops posting or commands stop responding, then in a few more seconds... disconnect.

At first glance, seems like a network problem, doesn't it? (Just to be sure: it disconnects and you have to restart, not crashes itself to the desktop, right?) If you haven't already, maybe reboot the modem as a kind of low-effort experiment?

Because the problem just started with the new release, might want to check-in with Firestorm Support (links just above) in case it could be a known problem.

[JeromFranzic]

Yep... looks like a FS problem. It happened to me a few minutes ago in Open Sim. Just sent a bug report via bugtracker, in Windows 11. Regular release 6.6.17 (with OS support), not the PBR viewer.

[AmeliaJ08]

2 hours ago, JeromFranzic said:

Yep... looks like a FS problem. It happened to me a few minutes ago in Open Sim. Just sent a bug report via bugtracker, in Windows 11. Regular release 6.6.17 (with OS support), not the PBR viewer.

I used to get quite a lot of sudden crash-to-desktop with bugsplat type crashes on OpenSim grids using Firestorm. I think in general OpenSim grids cause more of this kind of thing, it often seemed to be related to nearby regions (could replicate crashes by travelling close to region boundaries) but was the type of crash I have never seen replicated in Second Life.

 

[AmeliaJ08]

12 hours ago, Brises Ravenheart said:

Brand new Acer ROG laptop, I7  processor with Nvidea RTX 4060, runs fantastic with sometimes as high as 160 fps. new install of firestorm, and crashing many times per day. The only clue I have is text stops posting or commands stop responding, then in a few more seconds... disconnect. This was happening with my previous laptop as well, but I had managed to get it more under control, still happened though. Not sure if this is a firestorm issue of a Linden issue. Tried the scan from the previous post, and  "Windows resource protection did not find any integrity violations."

That definitely sounds like a network issue, not really a crash. I know people call a disconnection a crash though, it's just a terminology thing.

As far as network issues go poor WiFi signal strength can cause this type of disconnection, WiFi card driver (check for updates from manufacturer), poorly performing VPN connection is also a common cause. The teleport disconnection bug is still a thing in general too but you wouldn't normally experience that more than once or twice a day, anything more points to a different cause.

 

[Jessicathenobody]

Necrothread bump.

 

i'm currently struggling with similar issues to the OP, and i hate that it might be a hardware issue.  i've two modest older systems running 8gb RAM each and GT 635 cards, and i'm not really able to place one through its paces but the one i use for SL just keeps crashing, and often a relog gives me the cloud, contact list and inventory issues until i exit and do it again.  This is Firestorm and Alchemy, in Ubuntu and Windows.  The official SL viewer in Windows will give me the BSOD.  Sometimes in Linux the third party viewers simply crash, sometimes they freeze the system up.  i don't know.  i've run Memtest from Grub, and let the whole thing run though to end with no errors.  Got another GT 635 coming and have downloaded Uningine Superposition to try and force a non-SL-viewer crash.  This is getting very frustrating.

Edited March 15 by Jessicathenobody

[AmeliaJ08]

edit: grrr necrothread! I tried replying to OP from 2022 😂

 

 

Edited March 15 by AmeliaJ08

[AmeliaJ08]

3 hours ago, Jessicathenobody said:

'm currently struggling with similar issues to the OP, and i hate that it might be a hardware issue.  i've two modest older systems running 8gb RAM each and GT 635 cards, and i'm not really able to place one through its paces but the one i use for SL just keeps crashing, and often a relog gives me the cloud, contact list and inventory issues until i exit and do it again.  This is Firestorm and Alchemy, in Ubuntu and Windows.  The official SL viewer in Windows will give me the BSOD.  Sometimes in Linux the third party viewers simply crash, sometimes they freeze the system up.  i don't know.  i've run Memtest from Grub, and let the whole thing run though to end with no errors.  Got another GT 635 coming and have downloaded Uningine Superposition to try and force a non-SL-viewer crash.  This is getting very frustrating.

BSODs do often point to hardware issues yes but you should try a clean install of the Nvidia driver just in case. Use DDU (Display Driver Uninstaller) and follow the steps to completely wipe the driver first and then install the latest supported driver for the GT 635 and report back.

SL should work on that card, I have used it on an old system with a GT640 no issues other than terrible performance and if RAM is testing out fine it leaves little other than (hopefully) driver issues at fault.

If you are getting it in Linux as well though... video memory fault? overheating? hard to say. Can you install HWInfo in Windows and monitor the GPU sensors? it might be as simple as overheating, if the card has never been repasted for example it definitely needs it by now at its age and those GPUs were notoriously hot... given it is happening in Windows and Linux I would definitely look at disassembling the card, cleaning everything, repaste and replace the memory thermal pads if necessary.

Edited March 15 by AmeliaJ08

[Jessicathenobody]

Thanks for responding.  Heat has occurred to me, when i ran that benchmark it got up to maybe 80C.  Repasting hadn't occurred to me, though.  i should have some around somewhere.

That benchmark displays temperature as it runs; i was getting about 80c during the test.  Is that excessive?  i don't know, finding operating temp specs has been difficult.

 

6 hours ago, AmeliaJ08 said:

BSODs do often point to hardware issues yes but you should try a clean install of the Nvidia driver just in case. Use DDU (Display Driver Uninstaller) and follow the steps to completely wipe the driver first and then install the latest supported driver for the GT 635 and report back.

 

Except that earlier in this thread someone mentions issues after a driver update they were having issues...

On 2/23/2022 at 10:08 PM, Paulsian said:

Few months ago my official viewer was crashing like few times a week and then a few times a day, finally figured out that the nivida driver I updated from their site was causing my entire computer to crash not just the viewer. Did you recently have a video card driver update and it was nivida? just throwing that out there. Other than that nope never had random perpetual crashes with just the viewer before. 

To that end, i thought i'd go to an older driver.  i mean, the card's pretty ancient, how many times can a driver be updated?  

i was having issues trying to do it in Ubuntu- i did manage to back up some with the Windows driver, but the downloads from NVIDIA only go so far back, and it is still crashing Windows.  At this point i'll also mention that on both my desktops, running FS has meant occasional screen freezes that last maybe 20 seconds or so, and that's gone on for a long time now.  That's been with Ubuntu, Mint, MX Linux, and antiX.  And the same issue running Windows, just not as bad.  But then this started up with the crashes on the computer i now use for SL, and it's also the latest one operating systems and drivers have been installed on.  

i'm not super savvy with command-line stuff, and i was running into issues trying to roll back the driver in Ubuntu.  The whole reason i was running Ubuntu was out of desperation- that's the distro FS for Linux is developed on.  Normally i'm a MX gurl. So Ubuntu showed me the driver i wanted to try but would error out trying to install any but the latest, and trying to do it manually meant all sorts of things i'm not familiar with involving shutting down the entire graphical interface.  So... scrap Ubuntu.  i reinstalled my usual daily driver, MX Linux (Debian based distro) and when it came time to install the driver, i was able to easily install the older driver i have wanted to try, through the OS, without having to be too familiar with command line operation.

And i spent... probably three or four hours in world today, and so far, knock on wood textures, it's run superbly.  No freezes, no crashes.  Longer than it's been being between issues, so i'm hopeful that this was really a driver issue.  i'm tempted to run in Windows again and see if that still crashes or freezes, and then i'll really and truly think it may be solved.  We'll see.  Thanks for responding, i'll keep the thermal paste in mind if this doesn't work, and when that spare card comes in i may just go ahead and take it apart and repaste it, then install that card.  Thanks again and i'll update this thread after i've found if my driver change works or not in the long term.  

[AmeliaJ08]

6 hours ago, Jessicathenobody said:

Except that earlier in this thread someone mentions issues after a driver update they were having issues...

It's possible, exceedingly unlikely Nvidia have broken the OpenGL driver that much in an update though. To be honest I'm not sure there's ever much work done on the OpenGL side of things but it's a useful enough troubleshooting step to try.

80c is fine for that card, I have seen cards where the issue isn't core temperature but simply bad contact from the memory to the heatsink though. Usually one comes with the other but old/missing memory thermal pads can definitely cause some instability.

[Jessicathenobody]

Thanks for the info.  i have a spare card coming and might take it apart from the get go- i'd updated a now gone computer couple years back an think i have some pretty good thermal grease laying around.  

i didn't think 80c was high, while running Firestorm i pulled up the NVIDIA server app an the temp is running about 70, 71.  But if i understand you right there might be hotspots?

I'd thought using the older driver fixed the issue- well, hoped.  But after i wrote my previous post, i went in world today an started having problems again.  It is so odd how it comes an goes.  For the time being, my workaround seems to be using Singularity.  It's old, it's slow, it loads textures at a snail's pace, but no crashes.  

Worst case, i switch desktops around an see how the other one behaves.  A little bit of a pita but doable.  

[Emma Krokus]

35 minutes ago, Jessicathenobody said:

Thanks for the info.  i have a spare card coming and might take it apart from the get go- i'd updated a now gone computer couple years back an think i have some pretty good thermal grease laying around.  

i didn't think 80c was high, while running Firestorm i pulled up the NVIDIA server app an the temp is running about 70, 71.  But if i understand you right there might be hotspots?

I'd thought using the older driver fixed the issue- well, hoped.  But after i wrote my previous post, i went in world today an started having problems again.  It is so odd how it comes an goes.  For the time being, my workaround seems to be using Singularity.  It's old, it's slow, it loads textures at a snail's pace, but no crashes.  

Worst case, i switch desktops around an see how the other one behaves.  A little bit of a pita but doable.  

Perhaps while you're waiting to sort out Firestorm running, you might like to try Genesis Viewer. It's a fork of the Singularity viewer and I find that this release works well for me: Genesis (64 bit) (1.8.957) Oct  8 2023 15:13:58 (Channel: Genesis Release) 

It is one of the TPVs - here's a website link: Genesis Viewer

Edited March 16 by Emma Krokus

[AmeliaJ08]

56 minutes ago, Jessicathenobody said:

Thanks for the info.  i have a spare card coming and might take it apart from the get go- i'd updated a now gone computer couple years back an think i have some pretty good thermal grease laying around.  

i didn't think 80c was high, while running Firestorm i pulled up the NVIDIA server app an the temp is running about 70, 71.  But if i understand you right there might be hotspots?

I'd thought using the older driver fixed the issue- well, hoped.  But after i wrote my previous post, i went in world today an started having problems again.  It is so odd how it comes an goes.  For the time being, my workaround seems to be using Singularity.  It's old, it's slow, it loads textures at a snail's pace, but no crashes.  

Worst case, i switch desktops around an see how the other one behaves.  A little bit of a pita but doable.  

Is Singularity able to address all of the available VRAM? I can't remember. I know some viewers are stuck on the old 512MB limit.

It's definitely odd you would only be crashing in one particular viewer but crashes like this can sometimes be blamed on bad VRAM, if Firestorm is attempting to use more than Singularity and hitting a potentially bad chip maybe that is it? it seems unlikely but... who knows.

 

 

Edited March 16 by AmeliaJ08

[Jessicathenobody]

Well an update...

 

I never got my replacement video card.  And it's just been getting worse.  Happens in Linux, Windows, happens with Singularity just not quite as much.  Well other night I got to looking at the program- I run it from the terminal in Linux, and even though I don't know much, thought I'd see if anything stood out when it crash, and found reference to 'pipeline error'.  Looking that up, that means likely memory issues.  Even though as I think in this thread or another one, it seems as if it often takes two logins after restarting to get inventory and names and textures to load, as if it's a network issue.  But anyway, even though I've run more than one kind of memory test without issues, I pulled out my modules, sprayed some contact cleaner on the slots and reseated the sticks, and so far today I've not had any issues.  Just as likely as not as soon as I post this, I'll have them again, but maybe not.  

I can't remember if I've tried that Genesis viewer or not, but I think I did start to download it, or maybe I tried to run it and it also gave me issues.  But anyway looking at the program dialog in the terminal after it crashes, I'm leaning heavily towards memory issues with the system RAM., thanks all who've responded.

[Henri Beauchamp]

3 hours ago, Jessicathenobody said:

 never got my replacement video card.  And it's just been getting worse.  Happens in Linux, Windows, happens with Singularity just not quite as much.  Well other night I got to looking at the program- I run it from the terminal in Linux, and even though I don't know much, thought I'd see if anything stood out when it crash, and found reference to 'pipeline error'.  Looking that up, that means likely memory issues.  Even though as I think in this thread or another one, it seems as if it often takes two logins after restarting to get inventory and names and textures to load, as if it's a network issue.  But anyway, even though I've run more than one kind of memory test without issues, I pulled out my modules, sprayed some contact cleaner on the slots and reseated the sticks, and so far today I've not had any issues.  Just as likely as not as soon as I post this, I'll have them again, but maybe not.  

I can't remember if I've tried that Genesis viewer or not, but I think I did start to download it, or maybe I tried to run it and it also gave me issues.  But anyway looking at the program dialog in the terminal after it crashes, I'm leaning heavily towards memory issues with the system RAM., thanks all who've responded.

Since you are not adverse to the v1 UI (Singularity, Genesis), give a try at the Cool VL Viewer: it got better VRAM and RAM management algorithms that might help you to keep things in line (for VRAM do have look to the Preferences floater, Graphics tab, GPU/GL features sub-tab, and do read the tool tips for the settings there: they are worth a user manual).

[Jessicathenobody]

I'll take a look at it, thanks!

[Jessicathenobody]

Especially that the crashes have started again.  No pipeline errors this time, just illegal instructions.  Gonna reinstall an' also look into Cool VL an' whatever other 3rd party viewers look interesting.  I mean I could run Windows but I resent the resource overhead associated with that OS.  I get somewhat better frame rates an' response when i stick to The Penguin.