One of the things I continue to like Python for is what I call 'bangingrocks together': quickly programming relatively small but non-trivialthings, on the order of a few hundreds lines and anywhere from anafternoon to a few days of time. A representative example would be thebasic UDP request/response bandwidth tester that I recently wrote; itcame to a bit under three hundred lines of Python with some comments,and took me perhaps a day or two of time to write, tune, make more

Since very few real applications just blindly fire UDP packets at atarget, when you want to know what UDP bandwidth you can get you aregenerally interested in the speed at which you can have a UDP-basedconversation between two sides. In other words, you need a bidirectionaltester, not just the sort of simple unidirectional one you can use withTCP bandwidth testing.

(Of course, TCP is actually bidirectional too. It's just that theconversation inherent in TCP is already handled by your TCP stack.

The traditional way that Unix kernels boot is to perform some very basicsystem setup, declare the running code to be a kernel process with PID 1,initialize all of the various important subsystems (such as networking),initialize drivers, mount the root filesystem from whatever device youtold them was the root device, and then just exec() (from inside thekernel) /sbin/init . At least in the somewhat old days, Linux was noexception to this.

The problem for initrds is that the job

For my sins, I have spent a modest amount of time testing how well a fewswitches did port mirroring . Thedepressing result was that neither of the switches I tested had portmirroring that worked really well for our needs, but on the other handit has given me an appreciation for the issues that port mirroring canhave.

(We use port mirroring for simple network traffic volume monitoring,and we would like to be able to use it for network debugging; at thetime I was investigating if we could use

There are a number of people who will sell you rather expensive networktap boxes (eg here ). Since hearingabout them and discovering the actual prices, I've felt that trafficmonitoring switches with mirroring ports (despite the VLAN issue ) and dual-NIC PCs running bridging software madethem pointless except for people with very high end needs; they wereneat in a theoretical way, but not something we would ever need inpractice, since the alternatives were perfectly good.

(There is a lot of elaborate equipment

One of the attractions of ZFS pools is that they are prettyself-documenting. A ZFS pool automatically captures a great deal of thebasic information that you need to deal with it, such as the filesystemsit contains and NFS export settings. But this isn't quite all of theinformation that you need for long term management and for disasterrecovery, and in our new fileserver environment we've opted to duplicate some of the information outside of ZFS as well.

(Keeping track of filesystems, export permissions,

Our old SAN was set up in the traditional way: the SAN backend units didRAID-5 interally, and this RAID-5 space was carved up into LUNs and usedby the frontend fileservers. For natural reasons all of our fileservers wound up using LUNs from all of our SAN backends.This setup has low overhead , decent resilience,and decent performance.Our new fileservers are set up inan entirely different way, and among other things they use RAID-1instead of RAID-5. Although we were

Our new fileserver setup requiresthat all of the LUNs exported by the backend SAN be exactly the samesize, down to the block, so that we can always mirror two arbitrary LUNstogether. We wanted our basic LUN size to be around 200 to 250 GB, butthat's a large range of possible sizes, and it still left us to pick asensible exact block size.

Our first approach was to take our 750 GB Seagate SATA disks and dividethem up into three equal

To simplify, BitTorrent breaks files up into blocks, checksums eachblock, and then has all of the clients (the peers) swap blocks back andforth. The checksums for all the blocks are in the .torrent metainfofile that each client has, so you know when you get a bad copy of ablock (and as a side effect, a malicious client can't serve you trojandata).

(One might think that trojan data would be relatively harmless. ButBitTorrent is used to distribute things that