Sorry, I did not make myself clear; I have been using daemontools for a few years and I am quite familiar with softlimit.
On Fri, 07 Nov 2003 09:33:15 -0700, Mike Roest <[EMAIL PROTECTED]> wrote:
Henry Baragar wrote:My question was about the reason for your modification of the run script. If your reason is compelling for a large number of installations, then Andreas should consider modifying the run scripts in the distribution.
Gustav,
Since softlimit is not part of the binc distribution, how did it get into the script? What was your rationale for adding it? Should it be added to the distribution? If so, what is a good value for -m?
Softlimit is part of daemontools so it doesn't need to be added to the distribution I don't think. I'm guessing his rationale was to prevent resource exhaustion attacks. Softlimit will limit the available resources to the child program. The -m limit is something that needs to be set on a site to site basis as for different people it will need to be different values. I'm not sure of the internals of how binc deals with large messages I'm guessing dealing with larger messages shouldn't require higher resource limits but Andy should be able to verify this.
I agree that 30M seems quite high. Does a value less than 30M work for you? Should we get Andreas to investigate to see if binc has a memory hog (or leak) somewhere?
One of the previous posts has the softlimit set to 30 megs which seems quite high. As combined my binc processes for accessing 5 accounts are using less memory then that.
Henry
-- Henry Baragar Principal, Technical Architecture 416-453-5626 Instantiated Software Inc. http://www.instantiated.ca
