Thoughts on systems safety, software operations, and sociotechnical systems.
This was a fun talk. I saw a lot of nods in the audience when I mentioned things pertaining to social applications (unpredictable usage, etc.). A lot of folks ask questions about how we use ganglia at Flickr. A PDF of my slides are here. If anyone can tell me how to get Keynote2 slides […]
Read more Slides from ‘Capacity Planning for LAMP’ talk at MySQL Conf 2007
Caching systems are finite in size. So what happens when your cache is filled with objects ? No more objects ? Game over ? Hopefully, no. Most modern caches have some form of replacement or eviction policy. What means that based on some criteria, it’ll figure out what objects to throw out the window so […]
Looks like I’ll be talking about capacity planning. See you there!
I don’t think that these have been built yet by anyone, but I want these: 1. Swap Inspection Tool – a utility that can simply peer into swap space and tell me whatever it can about what is using it. PIDs, process names, anything. I know it’s just a snapshot, and it could all change […]
This looks promising and potentially wickedly awesome: http://www.linuxvirtualserver.org/software/tcpsp/ Has anyone played with this yet in a production environment ?
Ok, Geva Perry tagged me so here goes: 1. After graduating from Umass/Amherst with a Mechanical Engineering degree, I worked at the Volpe National Transportation Systems Center, doing side-impact crashworthiness research for the NHTSA. We ran transient dynamic finite-element analysis simulations using LS-DYNA and rigid-body simulations with MADYMO3D. This is where I learned UNIX, because […]
I hate building packages, but it’s a necessary part of the job sometimes. I have loved checkinstall for years, which basically replaces “make install” with something more like “make an rpm package for me”. But of course, when you have a simple tarball, and there’s no make install step to replace, then what ? One […]
1 – I’m glad that tcpdump has been the standard for capturing network traffic, cause I’ve learned a lot over the past 12 years using it. 2 – tcpdump tools suck. But I’ve been using pktstat recently, and it’s pretty damn good. ‘top’ for network, without all that stuff that ntop comes with.
when you see this: (loaded box, lots of network, 2.6.9.x kernel, etc.) Dec 21 04:05:58 hostname kernel: ————[ cut here ]———— Dec 21 04:05:58 hostname kernel: kernel BUG at net/ipv4/tcp_output.c:927! Dec 21 04:05:58 hostname kernel: invalid operand: 0000 [#1] Dec 21 04:05:58 hostname kernel: SMP do this: Add to sysctl.conf: net.ipv4.tcp_retrans_collapse = 0
So there’s lots of excitement around Varnish, which is a caching proxy that is built to be first and foremost a reverse-proxy, as opposed to squid, which does both forward and reverse. Acceleration (reverse-proxying) is obviously important to us at Flickr, as we use squid extensively.