3,226 reputation
420
bio website bo.lt
location SF, CA
age 34
visits member for 3 years, 4 months
seen Sep 12 at 6:30

I love uptime, beer and motorcycles.


Oct
6
answered How passively monitor for tcp packet loss? (Linux)
Oct
6
comment Easy way to “edit” the traffic coming from a tcp host (linux)
Are these HTTP requests?
Oct
5
comment Initializing squid stops
You should probably uncomment the cache_dir line.
Oct
5
comment TCP Sessions Hanging with Debian and iptables
Regarding the interface drops, I'm assuming the external interface is seeing traffic for MACs it isn't responsible for. Getting tcpdump of all traffic on there with -e would confirm that. Also if you switch physical interfaces on the router and do wget tests to the router is everything still ok? (verifying that both interfaces are fine)
Oct
5
revised TCP Sessions Hanging with Debian and iptables
deleted 56 characters in body
Oct
5
comment OpenVPN connection from within 2nd subnet in office?
What happens if you peer with 10.67... when you are on that network?
Oct
5
comment Redundancy for WordPress sites
If you want a really good answer you should probably provide some more data. How often does data change? How often do you need it to update on the slaves? What sort of availability are you looking for? 98% or 99.9999%? When you say MySQL are you already running master/master? Do you have monitoring setup to track when certain limits are about to be hit so you can respond before they impact the site?
Oct
5
comment Force Apache to kill its children after lost connection (“stop button” pressed)
Are there tons of CGIs like this? You could write a perl wrapper that you called from apache that watched and if the output was closed killed its exec of the actual CGI. I don't think Apache has something built-in to do this.
Oct
5
comment In ip_conntrack, established but “stuck” TCP connections with huge time to live
You would except that it looks like conntrack is based on packets seen not the actual kernel states (its seems like it was built with the intent of tracking firewalls that usually won't have that in a local table). So I think if the session was established and closes poorly (like the client just goes away). I think they will stay in this state in conntrack and it looks like you may end up having to bump the total # if you run into the limit in conntrack (which I would then say is using so little resources there its worth upping the limit if you are hitting it).
Oct
5
answered providing logs to the users
Oct
5
answered Duplicity fails on Too many open files?
Oct
5
revised Duplicity fails on Too many open files?
fixed formatting and cleared out unimportant data
Oct
5
answered In ip_conntrack, established but “stuck” TCP connections with huge time to live
Oct
5
answered Configuring postfix to handle email on multiple domains without being an open relay
Oct
5
comment iperf max udp multicast performance peaking at 10Mbit/s?
Multicast would be the first place to look, do you get the same limits for unicast TCP testing?
Oct
5
comment fio config to measure IOPS against provider SLA
I would hope they'd let you use more resources than their SLA if they aren't currently limited. When doing networking SLAs I've usually configured the QoS buckets such that users can use the entire pipe if they are the only users, but it starts limiting if there is contention. Is there some reason to think they haven't done something similar?
Oct
5
comment Recommended Linux distro for point-of-sale stations
I agree with @Ignacio Vazquez-Abrams. There are a couple other red flags in your question that make me think this is the wrong approach. I've done PoS architecture before and having each one handle their own packages or lack the ability to do nearly instant reinstalls will leave you in a world of hurt.
Oct
5
comment What is the benefit to having a network's HTTP traffic sent out on a different IP than HTTPS?
Do the requests have an X-Forwarded-For header? en.wikipedia.org/wiki/X-Forwarded-For
Oct
4
comment Prevent data corruption on ext4/Linux drive on power loss
Turning off write caching or running 'sync' on a regular basis would probably help in the short term.
Oct
4
answered /etc/passwd continually accessed