ressource problem on linux?

[info@xxxxxxxxxxxx]

Hi,

I'm using Tor 0.2.1.22 on Debian Lenny. I played a little bit with
Tor (so there are 10 instances of tor client running simultaneous). I
can see very often the following in log:

We tried for 15 seconds to connect to '111.222.333.444' using exit
'SoDesuKa'. Retrying on a new circuit.

This very often occurs on SoDesuKa and sometimes on some other nodes
to. There is also

Have tried resolving or connecting to address '111.222.333.444' at 3
different places. Giving up.

When enabling debug log, I can see

Feb 11 08:23:51.768 [debug]
connection_ap_handshake_rewrite_and_attach(): Client asked for
111.222.333.444:80
Feb 11 08:23:51.768 [debug] connection_ap_handshake_attach_circuit():
Attaching apconn to circ 3699 (stream 0 sec old).
Feb 11 08:23:51.768 [info] exit circ (length 3):
$B8E356A56EC7300CA87BE4FD0D8096EA6E9113E1(open) lanroamer(open)
CityTor(open)
Feb 11 08:23:51.768 [debug] link_apconn_to_circ(): attaching new conn
to circ. n_circ_id 3699.
Feb 11 08:23:51.768 [debug] connection_ap_handshake_send_begin():
Sending relay cell to begin stream 35585.
Feb 11 08:23:51.768 [debug] relay_send_command_from_edge():
delivering 1 cell forward.
Feb 11 08:23:51.768 [debug] relay_send_command_from_edge(): Sending a
RELAY_EARLY cell; 4 remaining.
Feb 11 08:23:51.768 [debug] circuit_package_relay_cell(): crypting a
layer of the relay cell.
Feb 11 08:23:51.768 [debug] circuit_package_relay_cell(): crypting a
layer of the relay cell.
Feb 11 08:23:51.768 [debug] circuit_package_relay_cell(): crypting a
layer of the relay cell.
Feb 11 08:23:51.768 [debug] append_cell_to_circuit_queue(): Made a
circuit active.
Feb 11 08:23:51.768 [debug] append_cell_to_circuit_queue(): Primed a
buffer.
Feb 11 08:23:51.768 [debug]
connection_or_flush_from_first_active_circuit(): Made a circuit
inactive.
Feb 11 08:23:51.768 [info] connection_ap_handshake_send_begin():
Address/port sent, ap socket 13, n_circ_id 3699
Feb 11 08:23:51.768 [info] connection_edge_process_inbuf(): data from
edge while in 'waiting for connect response' state. Leaving it on
buffer.
Feb 11 08:23:51.768 [debug] conn_write_callback(): socket 4 wants to
write.
Feb 11 08:23:51.768 [debug] flush_chunk_tls(): flushed 512 bytes, 0
ready to flush, 0 remain.
Feb 11 08:23:51.768 [debug] connection_handle_write(): After TLS
write of 512: 0 read, 586 written
Feb 11 08:23:52.100 [debug] global_write_bucket now 10485760.
Feb 11 08:23:53.032 [debug] conn_read_callback(): socket 4 wants to
read.
Feb 11 08:23:53.033 [debug] connection_read_to_buf(): 4: starting,
inbuf_datalen 0 (0 pending in tls object). at_most 16384.
Feb 11 08:23:53.033 [debug] connection_read_to_buf(): After TLS read
of 512: 586 read, 0 written
Feb 11 08:23:53.033 [debug] connection_or_process_cells_from_inbuf():
4: starting, inbuf_datalen 512 (0 pending in tls object).
Feb 11 08:23:53.033 [debug] relay_lookup_conn(): found conn for
stream 35585.
Feb 11 08:23:53.033 [debug] circuit_receive_relay_cell(): Sending to
origin.
Feb 11 08:23:53.033 [debug] connection_edge_process_relay_cell(): Now
seen 9 relay cells here.
Feb 11 08:23:53.033 [info] connection_ap_process_end_not_open():
Address '111.222.333.444' refused due to 'server out of resources'.
Considering retrying.
[...]
Feb 11 08:24:56.496 [info] connection_ap_process_end_not_open():
Address '111.222.333.444' refused due to 'misc error'. Considering
retrying.
Feb 11 08:24:56.496 [info] client_dns_incr_failures(): Address
111.222.333.444 now has 1 resolve failures.

This occurs only when running 10 instances, with 5 tor instances all
seems fine. So it seems to be a problem with file descriptors or tcp
ports? Can anyone help?


***********************************************************************
To unsubscribe, send an e-mail to majordomo@xxxxxxxxxxxxxx with
unsubscribe or-talk    in the body. http://archives.seul.org/or/talk/