Wandering Thoughts archives

2007-12-31: There are really two GPL v2 licenses
2007-12-30: The importance of killing processes in the right order
2007-12-29: SNI doesn't work in practice
2007-12-28: Please don't make me pick an account name
2007-12-27: Why I am not entirely fond of Solaris 10 x86's boot archive
2007-12-26: Two problems with Python's file iterators
2007-12-25: Process memory layout for 32-bit Linux programs
2007-12-24: The difference between shells that do job control and shells that don't
2007-12-23: Multihomed hosts and /etc/hosts
Shortening hostnames for fun and profit
2007-12-22: Getting a useful persistent VNC session
2007-12-21: A thought about Solaris 10 x86's boot process
2007-12-20: Virtualization does not eliminate security concerns
2007-12-19: Why setuid scripts are fundamentally a bad idea
How x86 Linux executes ELF programs
2007-12-17: What is a script language on Unix
2007-12-16: A thought on reading multiline records
2007-12-15: There are reasons for stupid anti-spam policies
2007-12-14: Check then use is a dangerous security pattern
2007-12-13: Doing one-shot booting with GRUB
2007-12-12: A more abstract view of the generalized open() issue
Implicit generalized open()'s are dangerous
2007-12-10: A surprise about Linux serial consoles
2007-12-09: I don't like smooth scrolling
2007-12-08: A depressing thing about phish spam
2007-12-07: Using Linux's magic SysRq feature
2007-12-06: Why large ISPs like SPF (the cynical view)
2007-12-05: Safely updating files that are read over NFS
2007-12-04: Dumb switches are now too smart for my good
2007-12-03: A comment spam precaution that didn't work out
2007-12-02: How my CGI to CGI/SCGI frontend works
2007-12-01: My expectations for responsible spider behavior
BitTorrent's protocol is not designed to hide
2007-11-29: Hashes are not complete protectors of privacy
2007-11-28: The problem the automounter was trying to solve
Another aphorism of system administration
2007-11-27: Taking advantage of polymorphic WSGI
2007-11-26: The two things I can mean by 'web spider'
2007-11-25: Why we may not be able to use ZFS
2007-11-24: Gotchas with dual-headed X with RandR on ATI cards
2007-11-23: My web spider technical requirements
2007-11-22: The different types of hash collisions
2007-11-21: Linux virtual terminals and the X server
2007-11-20: A lesson learned: Always upgrade Fedora with a respin CD
2007-11-19: What IPsec being mandatory in IPv6 really means
2007-11-18: I love Linux's serial console support
2007-11-17: Fighting spam always costs
2007-11-16: Improving your life with checklists
2007-11-15: Platform risk and platform (in)security
2007-11-14: What NAT is useful for
2007-11-13: Why vfork() got created (part 2)
2007-11-12: Matching words against a list in the Bourne Shell
2007-11-11: Why vfork() got created (part 1)
2007-11-10: The risks of spam filtering (part 2)
2007-11-09: An object oriented design mistake illustrated
2007-11-08: How Linux handles virtual memory overcommit
2007-11-07: Understanding the virtual memory overcommit issue
2007-11-06: A thought about competition between Red Hat Enterprise and CentOS
2007-11-05: Some notes on Solaris 10 U4 x86 as an iSCSI target
2007-11-04: Thinking through salts for passwords
2007-11-03: A thought about the speed of IPv6 deployment
2007-11-02: Note to self: check for gigabit Ethernet
2007-11-01: Dynamic rendering versus static rendering for websites
2007-10-31: What may be causing my random NumLock issues
Jumbo frames on gigabit Ethernet on Solaris 10 x86
2007-10-30: The problem with big systems
2007-10-29: My problem with learning new programming languages
2007-10-28: The inconvenience of some DWiki design choices
2007-10-27: Why I am not really interested in hearing blacklist appeals
2007-10-26: The Slashdot effect is not like regular load
2007-10-25: Long term storage management in the field
2007-10-24: The format of PTR records in Bind irritates me
2007-10-23: How we sized the overcommit ratio
2007-10-22: The dangerous appeal of the obvious
Vim options it turns out I want
2007-10-21: How I got a corrupted metadb replica that paniced Solaris 10 x86
2007-10-20: Why mail systems should not defer rejections to RCPT TO time
2007-10-19: Some notes on booting single user in x86 Solaris 10
2007-10-18: The Python marshal module versus the cPickle module
2007-10-17: Our experience with Linux's strict overcommit mode
2007-10-16: Our old mail system's configuration
2007-10-15: The arrogance of trying to design for long term storage management
2007-10-14: Why I think identity blurs into authority
2007-10-13: Weekly spam summary on October 13th, 2007
2007-10-12: Getting your networks to your racks
2007-10-11: A gotcha with command order in pipes
2007-10-10: How to properly look up hostnames from IP addresses
2007-10-09: A silly trick with X
2007-10-08: I have new system enthusiasm
2007-10-07: An improvement in my comment spam precautions
2007-10-06: Weekly spam summary on October 6th, 2007
2007-10-05: Why we don't use cable management arms
2007-10-04: The corollary to who actually benefits from bug reports
2007-10-03: A basic principle of system design
2007-10-02: A gotcha with 'bidirectional' pipes to commands
2007-10-01: How Exim determines the retry time for local deliveries
2007-09-30: Understanding Exim's weird way of doing retries
Weekly spam summary on September 29th, 2007
2007-09-29: The first rule of free email-based services
2007-09-28: Phase objects: simple decent error reporting for Python programs
2007-09-27: Fixing Ubuntu's ethN device names when you swap hardware
2007-09-26: Thinking about why Apache waits for CGIs to close standard output
2007-09-25: How to clear Solaris Volume Manager metadb replicas on Solaris 10 x86
2007-09-24: Assume the existence of folklore among your users
2007-09-23: Names are not cheap
2007-09-22: Weekly spam summary on September 22nd, 2007
2007-09-21: An interesting bind(2) failure
2007-09-20: Websites should not accept random parameters in requests
2007-09-19: The benefit of chronological blog navigation
2007-09-18: Linux NFS client kernel tunable settings
2007-09-17: How mmap(2) requires a unified buffer cache
2007-09-16: In praise of Python's Global Interpreter Lock
2007-09-15: Weekly spam summary on September 15th, 2007
2007-09-14: A thought on untyped languages
2007-09-13: Limiting a process's memory usage on Linux
2007-09-12: Mass scanning via POP3
2007-09-11: Why I dislike ATX power supplies
2007-09-10: A small drawback of 64-bit machines
2007-09-09: Rethinking my views of Fibrechannel
2007-09-08: Weekly spam summary on September 8th, 2007
2007-09-07: My view of what 'strongly typed' means
2007-09-06: When you don't want RAID-5
2007-09-05: Where to find specifications on HTTP POST behavior
Features that I wish ZFS had
2007-09-03: Why syndication feed readers ignore an entry's 'last updated' time
2007-09-02: I wish mailers had a real programming language
2007-09-01: Weekly spam summary on September 1st, 2007
2007-08-31: Partial entry syndication feeds and updated entries
2007-08-30: Using reverse proxies to unify web sites
2007-08-29: One limitation of internal charges
2007-08-28: A gotcha with Solaris Volume Manager metasets
2007-08-27: On the naming of machines (part 2)
2007-08-26: A limitation in Linux's policy based routing
2007-08-25: Weekly spam summary on August 25th, 2007
2007-08-24: Linux and accidentally multipathed disks
2007-08-23: The excessive cleverness of some people's reverse DNS
2007-08-22: Redirecting traffic to another machine with Linux's iptables
2007-08-21: The dilemma of website facing
2007-08-20: Recognizing phish spam from exceedingly RFC compliant mailers
2007-08-19: A realization about breadcrumbs
2007-08-18: Weekly spam summary on August 18th, 2007
2007-08-17: A thesis about language niches
2007-08-16: How not to set up your DNS (part 17)
How to tell a DNS no data reply from a lame delegation
2007-08-15: The problem that built C++ (a thesis)
2007-08-14: The benefits of growing your toolbox
2007-08-13: One problem with distributed identity systems
2007-08-12: Adventures in network design, illustrated by our new backbone connection
Weekly spam summary on August 11th, 2007
2007-08-11: On the performance of Python longs being used as bitmaps
2007-08-10: Clever large integers
2007-08-09: A gotcha with the format of dump archives
2007-08-08: Explaining some university staff peculiarities
2007-08-07: A surprise with the Provides header in RPM
2007-08-06: Implementing a preforking network server in Python
2007-08-05: Thinking about more text formatting for DWiki
2007-08-04: Weekly spam summary on August 4th, 2007
2007-08-03: The scope of shell history
Using WSGI for performance tuning
2007-08-02: Link: XML on the web summarized
2007-08-01: A ZFS-based fileserver design
2007-07-31: Consequences of allowing packages to quiz users at install time
2007-07-30: What we want out of our new fileserver design
2007-07-29: My standard for clear idioms in personal code
2007-07-28: Weekly spam summary on July 28th, 2007
2007-07-27: How big is the Slashdot effect?
2007-07-26: An unexpected performance stress test for DWiki
2007-07-25: Solaris Volume Manager and iSCSI: a problematic interaction
2007-07-24: Using iSCSI and AOE to create artificial disk errors
2007-07-23: An interesting issue when yum upgraded gaim
2007-07-22: Universities are not businesses: an implication
2007-07-21: Weekly spam summary on July 21st, 2007
2007-07-20: The downside of a unified buffer cache
2007-07-19: A safety tip: keep your different sorts of source trees separate
2007-07-18: Why I like Python's large integer support
2007-07-17: Random passwords are not necessarily good passwords
2007-07-16: Why SSL and name-based virtual hosts don't get along
2007-07-15: Problems I see with the ATA-over-Ethernet protocol
Weekly spam summary on July 14th, 2007
2007-07-14: Linux kernel asynchronous IO doesn't work on sockets
2007-07-13: You can't change a Python function's local variables from outside
2007-07-12: An interesting mistake with ZFS and iSCSI
2007-07-11: Why I wish Python had assignment in conditionals
2007-07-10: Thinking about the Python equivalents of C's !! double negation
2007-07-09: How many bits of information are in a password?
How not to set up your DNS (part 16)
2007-07-08: A suggestion for HMAC signature construction
2007-07-07: Weekly spam summary on July 7th, 2007
2007-07-06: What the flags on DNS query responses mean
How not to set up your DNS (part 15)
2007-07-05: What OpenID is good for
2007-07-04: What OpenID is (and is not)
Problems with EXA X acceleration on ATI cards in Fedora Core 6
2007-07-03: How not to set up your DNS (part 14)
2007-07-02: What the unified buffer cache is unifying
2007-07-01: The optimization rule for systems
2007-06-30: Weekly spam summary on June 30th, 2007
2007-06-29: The stupidity of being nickled and dimed by vendors
Why forwarding all email for users is dangerous
2007-06-28: Why I don't like inverted if conditionals
2007-06-27: Why you can't use object.__new__ on everything
2007-06-26: A small update on comment spammer behavior
2007-06-25: ZFS's issues with long term storage management
The advantage of a SAN
Painless long term storage management without disturbing users
2007-06-23: Weekly spam summary on June 23rd, 2007
2007-06-22: One reason why the Debian package format is not my favorite
A thought on web server capacity measurement
2007-06-20: More on slot wrapper objects
2007-06-19: Thinking about types of bugs (and static analysis)
2007-06-18: What are slot wrapper objects?
2007-06-17: A use for CSS adjacent selectors
2007-06-16: Weekly spam summary on June 16th, 2007
2007-06-15: Why I am in system administration instead of programming
2007-06-14: Getting source RPMs with yumdownloader (part 2)
2007-06-13: What NFS file-based locking problems can happen
2007-06-12: 'Argument list too long' is a misleading message
2007-06-11: Getting source RPMs with yumdownloader
2007-06-10: How NFS is unreliable for file-based locking
2007-06-09: Weekly spam summary on June 9th, 2007
2007-06-08: Some thoughts on Slashdot's moderation system
2007-06-07: Why I hate firewalls, especially stateful firewalls
2007-06-06: Why you want a filesystem consistency checker
2007-06-05: RPM's multi-architecture file ownership problem
2007-06-04: RAID-5 versus RAID-6
Why we need our SAN RAID controllers to support logical drives
2007-06-02: Weekly spam summary on June 2nd, 2007
2007-06-01: Why our Solaris fileservers still use the automounter
2007-05-31: Please test your error paths
2007-05-30: A gotcha with the automounter and loopback mounts
2007-05-29: On storing source code in some non-text format
2007-05-28: Why ZFS's data integrity is less important than Solaris's usability
2007-05-27: Paying for security exploits
2007-05-26: Weekly spam summary on May 26th, 2007
2007-05-25: If you want work done, you need to pay for it
The risks of spam filtering (part 1)
2007-05-24: Firefox preferences settings that I use
2007-05-23: The danger of a web server writeable document area
2007-05-22: A case for breaking the web server ownership guidelines
2007-05-21: Linux is a Unix
2007-05-20: Properties relevant to finding what class supplies a method
2007-05-19: Weekly spam summary on May 19th, 2007
2007-05-18: The difficulty of throwing things away
2007-05-17: Determining what superclass supplies a method in Python
2007-05-16: My rules of thumb for picking conference talks
2007-05-15: Sizing your swap space (part 2)
2007-05-14: Sizing your swap space (part 1)
2007-05-13: Unix folklore: your swap should be twice your RAM
2007-05-12: Weekly spam summary on May 12th, 2007
2007-05-11: Building your own kernel with the Fedora kernel configuration
2007-05-10: The other problem with network booting
2007-05-09: CSS separates layout from content less than you'd like
2007-05-08: Supporting the real world
2007-05-07: What computer security is
2007-05-06: What Linux bind mounts are really doing
2007-05-05: Weekly spam summary on May 5th, 2007
2007-05-04: A little twitch I have in X Windows
2007-05-03: A stupid switch configuration trick
2007-05-02: My view of Ubuntu
2007-05-01: Some good practices for web servers
2007-04-30: Being specific about where your systems are
2007-04-29: What matters about object oriented operating systems
Weekly spam summary on April 28th, 2007
2007-04-28: Why I no longer bother to complain to ISPs about spam
2007-04-27: My view of Solaris 10 summarized
2007-04-26: A clever way of doing IP address assignment
2007-04-25: What do Unix errno values mean?
2007-04-24: RPC is surprisingly expensive
2007-04-23: Extra security systems for Unix should be explicit, not implicit
2007-04-22: Why Red Hat 7.3 is (still) so present on servers
2007-04-21: Weekly spam summary on April 21st, 2007
2007-04-20: Why organizations buy software from commercial companies
2007-04-19: A thought about attitudes towards support requests
2007-04-18: An advantage of CSS layouts for accessibility
2007-04-17: Why the University of Toronto can't just use Google Mail
A disappointment with ZFS
2007-04-16: Using CSS instead of tables is still using a hack
2007-04-15: chkconfig --add considered misleading
2007-04-14: Weekly spam summary on April 14th, 2007
2007-04-13: A little gotcha about binding services to interfaces
2007-04-12: Why Evolution is not my favorite mail reader
2007-04-11: Users don't really benefit from filing bug reports
2007-04-10: A limitation of Debian's /etc/network/interfaces control file
2007-04-09: Why indirect xdm probably doesn't work on your Linux machine
2007-04-08: A limitation of Python properties
Weekly spam summary on April 7th, 2007
2007-04-07: An interesting observation about web cracker behavior
2007-04-06: Why network booting is not a good answer
2007-04-05: Wanted: remote controllable DVD drives
An irritation about rails
2007-04-04: A brief sad note about root passwords
Social problems are the real problems
2007-04-03: Why charging for email won't do what people want
2007-04-02: Some lighttpd bits
2007-04-01: An advantage of virtualization for license servers
Please leave IOError and OSError alone
Weekly spam summary on March 31st, 2007
2007-03-31: Created functions and exception stack backtraces
Microsoft has another problem
Link: you are what you code
A simple debugger feature that I would really like
2007-03-30: What an Ethernet splitter looks like
2007-03-29: Usability issues with blog URLs
2007-03-28: A surprise to remember about starting modern machines
Dual identity routing with Linux's policy based routing
2007-03-27: The VPN routing problem
Some problems with iSCSI on Solaris 10 (on x86)
2007-03-26: Quality Solaris 10 software:
2007-03-25: Fitts' Law and edge-flipping in window managers
2007-03-24: Weekly spam summary on March 24th, 2007
How comment spammers behave
Randomly engaging NumLock considered irritating
2007-03-23: Counterintuitive RAID read performance
2007-03-22: An irony of conditional GET for dynamic websites
Making user home directories on a stock Solaris machine
2007-03-21: Users are almost always right
2007-03-20: On educating users
2007-03-19: The three strata of virtualization
2007-03-18: Why Unix setuid is incompatible with real network filesystems
2007-03-17: Weekly spam summary on March 17th, 2007
2007-03-16: GRE is a translucent tunnel
Things I have learned while doing GRE tunnels on Linux
2007-03-15: An annoying limitation of Linux IPSec
2007-03-14: The problem of machine startup order dependencies
2007-03-13: Machine room archaeology
What I currently do to stop comment spam on WanderingThoughts
2007-03-12: New warning messages might as well be fatal errors
2007-03-11: I consider __dict__ an implementation detail
Weekly spam summary on March 10th, 2007
2007-03-10: What a sysadmin's machine should be able to do
2007-03-09: Why hashcash schemes for email will never be adopted
2007-03-08: A bad popup dialog
A belated set of more power consumption numbers
2007-03-07: A gripe about sun.com
2007-03-06: Why I don't like USB keyboards
2007-03-05: Some useful new Linux software RAID features
2007-03-04: Handling lines with something-separated fields for Python
Weekly spam summary on March 3rd, 2007
2007-03-03: On useful front-panel LEDs
2007-03-02: A story of network weirdness
2007-03-01: An irritating limitation of listening sockets
2007-02-28: Using Unix domain sockets from Python
2007-02-27: Things I did not know about until recently: Ethernet splitters
2007-02-26: Things I have learned about effective sysadmin meetings
2007-02-25: Ordered lists with named fields for Python
How CSLab currently does email anti-spam stuff
Weekly spam summary on February 24th, 2007
2007-02-24: Thesis: any server push technology inevitably breeds spam
The problem with browser minimum font size settings
Most world-editable wikis are doomed
Some things to remember when implementing __getitem__
2007-02-23: The downside of distinctive hostnames
2007-02-22: A simplified summary of Python's method resolution order
My zeroth law of compromised machines
A note about the ordering of mixin classes
2007-02-21: Fixing Python's string .join()
The quick overview of DiskSuite failover
2007-02-20: Getting around LiveJournal's new minimum page width
2007-02-19: What I currently know about Fibrechannel versus iSCSI versus AoE
Another aphorism of system administration
2007-02-18: Why we do NFS fileserving with a SAN
Weekly spam summary on February 17th, 2007
2007-02-17: Programming fun
2007-02-16: I hate hardware (Dell 2950 edition)
QOTD: There are three types of authentication
2007-02-15: Something all full-service backup systems should have
2007-02-14: RPM tricks for dealing with multiarch machines
2007-02-13: On Python's grammar
2007-02-12: Why thin clients are doomed (part 2)
2007-02-11: How to do locking in shell scripts
2007-02-10: Weekly spam summary on February 10th, 2007
A temptation with challenge/response anti-spam systems
Link: Why the ease of installing Java matters
2007-02-09: What System V did to the poor ln command
2007-02-08: The danger of validating your XHTML
2007-02-07: Why I think thin clients are doomed
2007-02-06: What the Solaris 8 nfs3_max_threads parameter probably controls
2007-02-05: An irritation with yum localupdate
Another small user interface suggestion
2007-02-04: A sysadmin twitch about dump
2007-02-03: Weekly spam summary on February 3rd, 2007
2007-02-02: A modern environment's need for broadband
Why don't SQL servers do a lot of caching?
2007-02-01: Transparent versus non-transparent caching
2007-01-31: The inherent fragility of complex systems (in system administration)
Why I am not fond of DHCP in lab environments
2007-01-29: A gotcha with <textarea>
How to have your web spider irritate me intensely (part 2)
2007-01-28: Why DWiki doesn't use fully REST-ful URLs
Weekly spam summary on January 27th, 2007
2007-01-27: Why I think that DNS whitelists are going to fail
2007-01-26: First impressions of pyOpenSSL
2007-01-25: A clever trick to deal with students powering off workstations
2007-01-24: An irritation with current GUI interfaces
2007-01-23: How to stop DiskSuite resyncing a mirror on Solaris 8
2007-01-22: Why host authentication has to check who the certificate belongs to
2007-01-21: Sometimes system administration requires a hacksaw
2007-01-20: Weekly spam summary on January 20th, 2007
Browsers are the wrong place to report HTML validation errors
2007-01-19: Link: Peter Gutmann on PKI
2007-01-18: Why I want direct certificate checking instead of having to rely on CAs
2007-01-17: A grump about the socket module's SSL support
2007-01-16: Why I don't have a GPG key
2007-01-15: Configuring VLANs on Fedora Core
2007-01-14: Wrapping exceptions versus propagating them untouched
An alarming reflex in my use of find
2007-01-13: Weekly spam summary on January 13th, 2007
2007-01-12: The easy way to get me to not comment on a weblog
2007-01-11: Xen versus VMware for me
An interesting Bourne shell limitation
2007-01-10: A gotcha with inetd/xinetd and popular UDP services
2007-01-09: Solving an automounter timeout problem with brute force
2007-01-08: What I really want from an automounter
2007-01-07: Another issue with C's volatile
2007-01-06: Weekly spam summary on Janury 6th, 2007
Fixing up .rpmnew files
2007-01-05: Henry Spencer on C's volatile
2007-01-04: The scars of my NPTL experience
2007-01-03: The problem with C's volatile qualifier
2007-01-02: My current views on webmail providers
2007-01-01: Solaris's impressive ABI compatibility
Link: Threads Cannot be Implemented as a Library

Page tools: See As Normal.
Login: Password:
Atom Syndication: Recent Pages, Recent Comments.

This dinky wiki is brought to you by the Insane Hackers Guild, Python sub-branch.