Cool VL Viewer forum

Hi Henri,

you wrote:

The hardware is well-monitored, and temperatures are well within operating parameters (my system has more fans than George Clooney ). I checked last night's messages with no find whatsoever. I will give it a try at Calintreau later, although I don't think anything will turn up.

Again, the system should be well-maintained -- X.Org X Server 1.7.6 with 260.19.26, so no immediate explanation here either:

Guess I have to bite the bullet for now and see if I can get some decent crash information out of this beast.

Love,
Lia

To be more specific, what you must look for are messages starting with "NVRM: Xid": they are the sign that something went wrong on the hardware side, usually a timeout (which results in a system "hiccup" or hang, and sometimes right out a crash). Note that I had once a dying 8800GT (a no-brand card: I'll never buy again such cards and will stick will well known manufacturers from now on) that started to show bus errors and timeouts even while at normal GPU temperature and at stock speed...

Note that I had to delay my migration from Mandriva 2009.0 to 2010.0 because the new Xorg server (v1.4 in 2009.0 and v1.6 in 2010.0) was not happy with the Nvidia drivers (the crashes I mentioned): since 260.19.21 is out, I never encountered again that crash and can now enjoy using Mandriva 2010.0.

However, I updated today to 260.19.26, and I just got an error that caused a graphic slow down (not even with the viewer open) and that could be related with this new driver version:
Thing is, IRQ 16 is used by the Nvidia driver, so the automatic disabling of IRQ 16 is a very bad idea... I added the "irqnodebug" boot option (which would prevent automatic disabling on spurious IRQs receipt) and will see what happens, but a downgrade to 260.19.21 might be in order...

To summarize: the very latest version of a software is not always the one that works best for you...

Same thing here.

And I noticed that if the viewer freezes completely, which it does from time to time (freezes, not crashes) and I check the running processes in Task-Manager it shows a whole bunch of running so called SLPlugin.exe show up there. Once I ended those manually the freezing is gone and the viewer works just fine again. (It does somehow affect my VLC media player though it seems.)

Anyways, I'm going to check if I might miss the latest driver for my Nvidia card...
Will report back afterwards.

@Vilandra

It would be interesting to know what media are attached to all those SLPlugin.exe processes: there should be at most a few such processes (6 or so max, when the search floater is open, a profile floater open and when a media is playing, all at the same time), and if you get many open permanently it shows that some somehow got stuck (and will of course cause slow downs)... Like always, I do need logs and precise repro steps (what region and parcel did you get all those media) to be able to investigate further, since I don't see such problems here...

You could also disable media playing and see how it fares.

Hello Henri

Those SLPlugin.exes didn't show up so far again. I have to say I tried some stuff. Installed the latest Snowglobe based Cool VL etc. With this one the SLPlugins seemd to have vanished (I remember the viewer giving me notices about the multimedia plugins when I shut them off manually the first time, which was a couple of days ago.).

Anyways, after updating my Nvidia drivers I started listening to my poor computer and everytime it freezed there was a real terrible scratching noise. So today I've been a bit busy. After a complete uninstall (including all possible left overs) I noticed some of the texturecache folders (They weren't there with 1.23, were they?) screwing things up a bit. I still can't delete those by getting some weird CRC error messages. So in the end I simple renamed the folder and installed Snowglobe and Cool VL brand new. This way I could make sure it would end up in brand new folder. So far it works and the freezing didn't occur again. Seems my poor little fellow harddrive is close to it's end. *sighs* Maybe I should write a wish-letter to dear Santa

In case of any news I see what I can do to get the crash logs.

Keep you posted,
Vila

that sounds like your computer is having hardware trouble... just my 2 cents.

Yes, it's definitely a sign of a soon to be dead drive: do backup all your data and be prepared for a dead system in the coming days or weeks... Note that dying hard drives can also be the cause for system 'hiccups' and crashes: when the drive notices it can't read a sector, it retries it a dozen of times (which may be the cause for a first 'hiccup'), then, if it succeeds it might (depending on how the firmware of the drive is coded) try and reallocate the bad sector (it marks the physical sector as bad then uses one of its few spare sectors as a replacement, setting the logical sector to the spare physical sector): this can also be the cause of a second (or of a longer) hiccup. Finally, if it fails to read the sector at all, it reports the error to the system, and depending on the reason why the system was trying to read that sector, it may crash a process (in your case the SLPlugin.exe, most probably), report a read/write/CRC error, or just stupidly crash...

Your hard drive might still be under warranty: Western Digital got a 3 years warranty (5 for "Black" editions), Seagate offers 5 years... I myself had to return a couple of drives: one 15000 hours, 2 years old Seagate drive that was showing reallocated sectors and was causing system hiccups before it finally reported read errors, and one Western Digital among three I bought as replacements for the Seagate drives of a RAID 5 array (the other 2 drives were also 15000 hours old and I could not trust them after the first one died) and that was "dead" (permanent read errors on one sector) on arrival... These days, drives are pretty fragile compared to older, lower capacity drives (got a couple of such old drives that are still healthy after over 50000 hours of 24/7 operation). My faulty drives were both replaced under warranty, and WD even allowed an advance return procedure (they send you a new drive, then once you got it, you return your faulty drive).

My advice is therefore to check for the warranty status of your drive on the manufacturer's website, and to download a diagnosis tool from them (try Seatool from Seagate: it also works with other brands). Before running the diagnosis tool, do make sure to backup all your data however (the tool will put the drive under stress, and some tests you might want to run such as "zero drive" wipe all the data).

"backup your data" these days is easier said then done... when a "personal computer" has more storage space than your average bank datacenter 10 years ago.
How do you reliably backup two terabyte in a timely manner without spending tons of money for new hardware?

That's why I'm using RAID 5 arrays... When one drive fails, you still have all your data intact and you need only to replace one drive to restore the full reliability and readiness of your array... However, *important* data must still be backed up (and it rarely represents 2 Tb of data) , since you may still get your live data erased by mistake or because or a software bug... This backup may perfectly live on another partition of your RAID array (one of those partitions you only mount and use for backups and normally keep untouched), if you are confident enough in your array (note that hardware RAID should not give you such confidence since while a software RAID array may be ported to a different machine with different hardware, a hardware RAID is only as reliable as the hardware is: if you loose your hardware RAID controller, you may loose all your data if you can't get a replacement for that controller).

raid5. at home. i am SO not going anywhere near your computer without earmuffs. XD