Strange Integer error

[mescal1n]

Hi ...

I'm running a very good working ioftpd v7.0.3 system on my own PC, but wanted to help a good friend setting up a fresh system with latest ioftpd etc. on a WinXP-system.

So I've downloaded ioftpd v7.5.9 and installed latest ioninja and nxtools too. Everything is working well so far. Except uploading dirs with more then ~260mb. Till < 260mb dirs, all is fine. After uploading the next file which hits the ~260mb mark i get this strange error in the systemerror.log:

01-19-2011 20:03:16 "..\scripts\ioNiNJA\ioNiNJA.itcl" terminated abnormally
--- ErrorInfo ---
integer value too large to represent
while executing
"lsort -decreasing -integer -index 2 "$total_racers""
(procedure "outstats" line 68)
invoked from within
"outstats 0 0 0 1"
(procedure "::ioNiNJA::ZipScript::postdel" line 50)
invoked from within
"::ioNiNJA::ZipScript::postdel "
("DELE" arm line 1)
invoked from within
"switch -- [lindex $args 0] {
SYMCLEAN {
set temp [userfile open


I have put a iputs $total_racers one line before lsort -decreasing -integer -index 2 "$total_racers" in the ioninja.itcl to see $total_racers in ftp output which showed me:

[2] 226-{mescal1n 4085577465 27237051271 363160679741 2421071198273}

Can someone imagine what could be wrong? I mean where from come these large numbers? Wat could that be? :confused:

Thanks for your help.

[Yil]

That's a new one to me. I know that combo of script+server has been working fine on lots of systems with large GB dirs so I'm guessing it's something really strange...

Here's a couple of things to double check. Is TCL installed anywhere else on the machine? If you check the user's stats via site uinfo do they seem correct and not some huge numbers? If you've customized the setup, perhaps try grabbing a new copy from the zip and trying a minimum setup just to see if the zipscript works without nxtools and any customizations? If you copy the setup from his machine to yours do you get the same problem?

[mescal1n]

Is TCL installed anywhere else on the machine?

No.

If you check the user's stats via site uinfo do they seem correct and not some huge numbers?

No, all fine. But I checked "racestats" ... let me show you:


.-----------------====--------------------------== ioNiNJA v0.8 ===--.
| |
|-=[ R A C E N F O ]=---------------. |
| `-----------------====----------|
| Release: testdir |
| RaceNfo: 128MB in 9F @ 1.3MB/s in 1m 40s |
| Fastest: mescal1n/ @ 2.0MB/s |
| Slowest: mescal1n/ @ 2.0MB/s |
| |
|-=[ U S E R T O P ]=---------------. |
| `-----------------====----------|
| 1. mescal1n 10.8GB 2801 1% 246262.7MB |
|- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - |
| 1. Total 128MB 9F 100% 1.3MB/s |
| |
|-=[ G R O U P T O P ]=-------------. |
| `--------------===--------------|
| 1. 1036948 10.8GB 2801 % 246262.7MB |
|- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - |
| 1. Total 128MB 9F 100% 1.3MB/s |
`-----================---------------------------==------------------'

The user and group stats are totally wrong, but the total on user and group are correct.
Also group name should be STAFF and not 1036948. :confused:

If you've customized the setup, perhaps try grabbing a new copy from the zip and trying a minimum setup just to see if the zipscript works without nxtools and any customizations?

Tested too, but same effect.

If you copy the setup from his machine to yours do you get the same problem?
Will check this out now ...

[mescal1n]

If you copy the setup from his machine to yours do you get the same problem

I've tried that, just changed network settings in ioftpd.ini.

Same error.

fail :(

[Yil]

Very weird... If it's repeatable and transferable across machines then something is clearly wrong, but in that case lots of people should be having the issue.

Your older setup works fine, but a copy of his new setup to your machine fails. Perhaps you could try grabbing a NEW copy of the server and ninja and create a new version on your machine. If that fails as well then something is really odd and we need to start looking into other installed software on the machine that might be causing issues because clearly that should work. Because it failed on 2 different machines we only have to find the stuff in common between them and that isn't stock windows stuff.

You can use the 'site crashnow' command to force the server to generate debug files (and exit). You can examine the CrashLog.txt file to see what files were loaded into the server and figure out which ones aren't MS .dll's to see if something stands out. I've had problems with many injected dlls in the past but these have usually caused crashes or network issues and not something as subtle as what you are seeing.

As a last step you can package up the crash info and zip up the config and send it to me just to see if it works for me or if I can spot something. I'll PM you the debug FTP to upload it to.