Nothing in log, and ioFTPD-Watch.exe is running using 996K ram. I did however notice something I haven't seen before when examining event prior to hangup:
sitebot only reported newdir creation on three different releases, as if zipscript stopped working entirely. Three minutes earlier it put out full output with a working zs. Type of files was excactly the same in both before and after. In addition to this, when examining logs, I noticed that io had stopped accepting new connections 20 MINUTES before. (I noticed this due to a user hammering every 20s but getting login refused in log) This particular user had been hammering for a full hour already, and is also the last entry in ioftpd.log so it's unlikely that he stopped hammering at that particular point.
This leads me to believe that already connected users aren't very much affected by the lockup bug, and that it is possibly connected with the execution of zipscript.
Using ioFTPD 7.3.2 - dZSbot - ioninja customized PZS-NG 21.1
|