Fatal error [023.024] RLEUnpack failed?

[furballexpress]

OK, this is a machine that's run the client through one generation with no errors. It's a duallie running one DF client as a service on NT4 SP6a.

========================[ Jul 16, 2003 3:53 PM ]========================
FATAL ERROR: [023.024] {trajtools.c, line 2507} RLEUnPack failed, size=3, should=400*400 - likely this is caused by overclocked or faulty RAM chips, please test your RAM

========================[ Jul 18, 2003 10:21 AM ]========================
FATAL ERROR: [023.024] {trajtools.c, line 2507} RLEUnPack failed, size=3, should=400*400 - likely this is caused by overclocked or faulty RAM chips, please test your RAM

========================[ Jul 18, 2003 10:22 AM ]========================
FATAL ERROR: [023.024] {trajtools.c, line 2507} RLEUnPack failed, size=3, should=400*400 - likely this is caused by overclocked or faulty RAM chips, please test your RAM

========================[ Jul 18, 2003 10:23 AM ]========================
FATAL ERROR: [023.024] {trajtools.c, line 2507} RLEUnPack failed, size=3, should=400*400 - likely this is caused by overclocked or faulty RAM chips, please test your RAM

========================[ Jul 18, 2003 10:23 AM ]========================
FATAL ERROR: [023.024] {trajtools.c, line 2507} RLEUnPack failed, size=3, should=400*400 - likely this is caused by overclocked or faulty RAM chips, please test your RAM


As this is a dev Compaq ML530 for testing, I DO NOT have it overclocked. Nothing else on the box fails/"abends"/errors with any memory error at all, and a recent mem read/write test passed with flying colors. The DF client has not failed before now either. The client is sneakernetted, but I move all files to a backup dir, sneaker, ftp u/l, sneaker back the u/l'ed dir, and delete old log.bz2, min.val and .val files as needed, and, as I said, this client has never had a problem, nor has the box it's on. Oh - and this is the original Phase2 client, not the patched one (notice I said it runs as a service above ;) ).

Hey! I think this might be a new one! w00t!

[Stardragon]

It seems to be a very isolated error - it's only occured on one machine and no one else has reported anything similar in the past month.

If anyone has had a similar error, please post to this thread.

Otherwise, I believe it's exactly what it says - bad/corrupted RAM. Try running some more tests or running on a different box to test.

[furballexpress]

Well....interesting.

The exact same thing happened to another ML530 i've been stealing cycles from. Similarly configured machine, this client isn't as old and has yet to finish a generation, but the error is exactly the same, and occurs every minute or so.

Stardragon - As much as I'd like to believe this is my hardware, I have to say both of these boxen have been running another application without fail or reboot for over a month. To be fair, I did reboot both machines and tried again before posting here...same result.

I *really* don't think it's the RAM, despite this problem's isolation to few systems. Could it be something to do with the interaction between the client and NT4 SP6a? Hotfix missing or something? I'd believe something like that before I'd believe it's the RAM. The read/write test performed does sequential, random and continuous read/writes to all memory registers (which took a while on the 3GB RAM in these boxen), and there were no errors found. Nothing on either box is overclocked or out of hardware spec one iota. Sorry....

[Brian the Fist]

This error refers to your protein*.trj file being corrupt. If you would like you can e-mail this file from the damaged machine to trades@mshri.on.ca and we can verify what is wrong with the file for you.