Search results

Nonsense. tmpfs on *NIX based servers uses both VFS and swap as a backend, and even if you paged it to disk RTT socket latency is STILL going to be an order of magnitude issue. Paging to disk is a good idea on already compiled opcode, keeping everything in memory is a pessimistic optimization...

APC is a memory hog, and doesn't page compiled opcode to disk. I recommend against it.

- Lots of memory for filesystem buffer cache.  - Depending on your 'linux kernel version', toggle kernel.vsyscall64 to 2 to give a little bit of a speedup on gettimeofday()  - Use a 32bit version of *NIX to save memory, as 64bit has swollen pointers / alignment padding issues that consumes...

Another useless benchmark. Apache with MPM Prefork? Are you fucking kidding me? Of course other webservers are going to win, vfork/fork() is expensive! "The static file was 100 bytes. We used such a small file to avoid saturating the network connection" 100 bytes? Well, litespeed does some...

Why do people care about this stuff? ioping is using a syscall to measure latency, and you cannot get accurate measurements from userland due to context switching. Not only that, a majority of the 'latency' comes from the controller itself which is far, far away behind some controller.

FreeBSD. 

Sar is useless if you want to pinpoint what is causing excessive iowait. 

My security tips: - If possible, avoid using any windows based browsers. Use vmware or something with IOMMU enabled in the BIOS, then do your web browsing etc securely within a virtual machine.  - Don't use java (period) - Don't use flash/shockwave (period) - Mangle your webbrowser useragent...

AFAIK they're running an XE version that I placed on a majority of the switches due to a no ip redirects + CEF bug I discovered where the switches would deadlock (the only time TAC came in handy) That specific 3550 was provisioned years ago when I was still an employee. Work pretty good if you...

When I spoke to you on the phone years ago you sounded drunk when I asked for the status of a customer provision. 

No. I wasn't /that/ stupid. gary@exar-kun:~$ DOMAIN=colocrossing.com ; for i in `whois $DOMAIN | grep "^Name Server:" | awk '{print $3}'` ; do dig +short www.google.com @$i ; done

Actually, I take that back, I wrote the network specs and what we used at the time. Everything else is marking fluff.. 

I wrote the above years ago when we first turned up BGP and it's all correct. 

Thanks for noticing that. I've adjusted it, the original version had some custom things I could not publish due to company policy so I had to replace some things. (It was also written in C)

To use perl loadlogger.pl &  (or you could run it in a screen)

Instead of using usually bizarre RRDs to see load spikes on customer machines and the likes, and scraping through piles of logs that show nothing, here's a small script that I quickly wrote that is used in a production environment to log excessive load (mostly ps data for tracking down runaway...

Sounds like a bunch of kids with PHP bots that do socket() commands as legitimate users. 

People still use the http keepalive dos tools? That's so.. 2006'ish. I have some penetration testing things I wrote about 6 years ago, one of them is a neat SNMP reflector that can generated about 10gB/s of traffic due to spoofed UDP+SNMP OID requests..  DDoS protection services are a total...

The phone-home blocks on TOR were not the FBI, they were a /16 owned by the NSA. 

Blocking ICMP completely is one the more silly ones I used to do, until I realize PMTU and others require working ICMP.