Hello all, [note: I'm posting this to NetBSD's pkgsrc-user mailing list, as well as the TOR talk mailing list] I'm trying to run tor v0.1.1.22 on NetBSD/amd64 3.0. I'm running it as an exit node. It does seem to run pretty fine, but only for a while. I noticed that it does consume *vast* amounts of RAM, and after a while it has consumed so much that it starts eating off the swap, and finally I get messages which state that TOR processes are being killed off for eating up the remaining swap. This is an except from "top" (yes, I am aware of that "top" isn't optimal for memory related detective work). PID USERNAME PRI NICE SIZE RES STATE TIME WCPU CPU COMMAND 1535 tor 2 0 38M 23M netio 0:00 0.00% 0.00% tor 1202 tor 2 0 38M 23M netio 0:00 0.00% 0.00% tor 709 tor 2 0 38M 23M netio 0:00 0.00% 0.00% tor 1823 tor 2 0 38M 23M netio 0:00 0.00% 0.00% tor 1126 tor 2 0 38M 22M netio 0:00 0.00% 0.00% tor 1069 tor 2 0 38M 22M netio 0:00 0.00% 0.00% tor 1716 tor 2 0 38M 22M netio 0:00 0.00% 0.00% tor 1688 tor 2 0 38M 22M netio 0:00 0.00% 0.00% tor 1198 tor 2 0 38M 22M netio 0:00 0.00% 0.00% tor 1150 tor 2 0 38M 22M netio 0:00 0.00% 0.00% tor I'm not really sure how an application which relays TCP traffic (albeit encrypting/decrypting it) can consume this amount of RAM. Is it really normal? What's worse... A few times NetBSD has hung, *completely* (requiring hard reset), when TOR has consumed all memory. I have never has NetBSD crash hard on me before as far as I can remember. Since I started running TOR a few days ago, I have had several total crashes. Is anyone else having problems with tor and memory consumption? On NetBSD? -- Kind regards, Jan Danielsson
Attachment:
signature.asc
Description: OpenPGP digital signature