socat dying intermittently

[rabbas]

Hello,
I have an asus RT-AC68U. I have installed the latest version of Asus Merlin. I have a socat process that redirects packets via an ssh tunnel to a server in the cloud. These are started via post-mount scripts.

I have a local ubuntu server doing the same thing and an openwrt (running as service).

Everthing works well but once in a whilte I have socat shutting down with the following error: "socat[1250] E fork(): Cannot allocate memory"

From the error it seems that there is no memory for a new socat fork. My typical memory usage is as follows
admin@RT-AC68U-BA30:/jffs/scripts# free -mh
total used free shared buffers cached
Mem: 255700 73968 181732 0 1232 13100
-/+ buffers/cache: 59636 196064
Swap: 0 0 0

I am not sure if this specifically with socat (as in memory leak) or it is general for my device.

Has anyone expericed this? Should I add swap?

Any advise would be greatly appreciated.

Raza

 

[Jack Yaz]

Yes add swap. While you won't see it used, it's known that a 2GB or more swap file alleviate cannot fork errors

 

[rabbas]

Thx Jack. I will give that a shot. For now I have turned off dlna and samba and to add to that I will activate swap.

 

[rabbas]

Following up. Adding swap certainly helped because it reduced the frequency with which the problem occurred. However it does occur but less often. For socat I am udp:recvfrom and this does not seem to support max children to control number of forks. I tried with udp:listen but that does not seem to work because it does not reply. Is there a way around this issue so that I can either control number of forks or an alternative to passing udp dns packets thru ssh tunnel.