Sudden string of crashes after days of bugginess

[FunkySwerve]

Sudden string of crashes after days of bugginess

Ok, this is a bit drawn out so please bear with me. On August 30th I started getting bizarre complaints about gear malfunctions. At the time, I thought, no big deal, players will be players. "yesterday i noticed with valcano helm on i was being knocked down in the tower from the humans with sonic swords. today a party member of mine was complaining he had the sage ring on and he kept getting confused in the tower from the dog creatures." etc etc. But then I started getting a few more such reports, and not from your typical clueless noob swearing up and down hes found a bug, but from players who've been on the server a loooong time. Then, a day or two later, the meteor swarm spell stopped working. It began to send the player the message, "there are no remaining uses of this spell" no matter how many they had memorized. I thought 'odd, I haven't altered anything with that spell in weeks'. So I tried running it on my machine, and it works fine. So I do an update to our professional host's version, figuring it got corrupted enroute, but it still sends the same odd message. And then today, I logged into the server today to hear that it'd been crashing sporadilcally all day long, and that NWNX is no longer restarting it as it was as recently as yesterday. Almost immediately, a player messages me saying that the server seems to crash whenever he tries to login characters from his other gamespy account. I said, 'huh, that makes no sense, since youre using identicle versions, right?' He was. I had him test it by trying to login a character from the other account and *boom* down goes the server. He said that it was locking up before displaying any characters, and crashing everytime. So we get the server back up, and 15 minutes later it crashes again. I thought it was oddd at the time, especially since NWNX didn't bring it back up. Tried to restart it 3 more times and each time it crashewd as players (not the one already discussed, he agreed not to login that account until we figured out the issue) attempted to login. We use a remote program using ssh (secure shell) to control the server, and each time I logged into restart the server the top command revealed that it was still 'running' at 99.9% cpu ('top' is a linux command displaying top cpu uses, for you windows people). At this point, we gave up, fearing massive file corruption or god knows what on our host's machine. When I emailed the host about these bizarre incidenhts, his first suggestion was:
"The problems that you are describing sound like known issues that I've
heard of before. My first recommendation is to check the message
boards and other support sites for the hak packs and other add-on
systems that you are using."
We are running CEP 1.52, NWN 1.66, NWNX2 for Linux (newest release, on the d/l's page on their forums), and NWNX-Letoscript 23-5. My host felt that HDD or RAM problems were NOT likely to be the cause, though I cannot swallow the proposition that these events are not causally linked in some fashion. As it stands, we are unable to run the server until we can determine the problem, so I am begging ANYONE who has any idea what the problem might be to shed some light here.
Thank you for your time,
FunkySwerve

[Acrodania]

The server crashing as players login or transition are classic signs of corrupted items in their inventory. It seems like a definate thing to check also because of the other oddities your players are experiencing....

If these issues just suddenly started and can't be traced to a module update I would view the server's memory or temperatures with suspicion. I have seen several machines with NWN issues that turned out to be memory even though they ran fine with any other software; only a hardware memory checker could verify the problem. If there is memory corruption (caused by memory errors, symptom == oddities when running) when a player logs out it can corrupt items or the bic.

Heat, caused by weak fans, should be easy to verify....

[Parsec]

Hmm leto 23-5 might be part of the problem. I run 23-4 here and noted on the letoscript forums that 23-5 had a few bad bugs so never updated to it.