My Mandriva 2006.0 install on my T40 Thinkpad laptop now has 30 days of uptime, yay!
20:05:51 up 30 days, 7:11, 1 user, load average: 0.47, 0.78, 4.51
This is a personal record for my laptop, and a nice state to be in. It also means that software suspend has survived ~ 60 suspends in a row, given that I average 2 a day in moving my laptop around.
While Linux still has plenty of issues in a Laptop environment, I have to say that I’m quite pleased on the state that I’m functioning in now. 🙂
Funny and quick. My laptop has 7 days uptime (my month of uptime was killed by iwpriv ath0 mode 2 while doing wpa_supplicant things. 3rd time I’ve done that, 3rd time I’ve gotten killed on it).
Even though wipes out much of my state, it doesn’t actually for my 2 most used applications, Firefox and XEmacs. Firefox has a Session Saver plugin, which is pretty handy. XEmacs just keeps track of all open buffers at all times, so an application restart brings up all the previous buffers. After many weeks of not closing buffers in XEmacs, I decided to just drop some so my Ctrl-` key cycles through a more reasonable number of tabs. It was sort of funny to walk through and kill buffers off one at a time, as it was almost a complete history of the last 2 weeks. 2 code reviews, an article that I last was writing a month ago, many todo lists, and 3 or 4 personal projects I’ve been hacking on. 🙂
As soon as I get around to writing my X-Chat plugin, I’ll have the last major daily application stateful enough to be useful.
One of the reasons I really like Mandriva Linux is that it sets up a lot of Linux applications very nicely out of the box. One of the stellar examples is the way that it sets up CUPS for printing. For instance, when I roam between, home, work, and space, the moment I attach to the network, I knows which printers are there, and those are the ones it tries to print to. I’ve got CUPS servers in all 3 places that server up a sane default, so I never have to pick a printer, and lpr does it all for me.
One of my nits with Mandiva, is I often can’t figure out exactly how it does these things. I figured out that it was using slp for subnet based broadcast (all cups servers with Broadcast enabled send these multicast packets around periodically to self configure with each other), which works fine at home, and at work if I’m wired, but doesn’t solve the whole story. The last bit of the puzzle is the use of BrowsePoll in your /etc/cups/cupsd.conf.
BrowsePoll specifies an IP:PORT that cups should try to poll for printers at the same interval as it is sending these slp packets on the network. Adding a couple of entries there, and I never have to think about printer configuration again. 🙂
After the last update to the madwifi-ng driver core Revision: 1396, the driver no longer causes a kernel panic when used in conjuntion with ipsec. This is great news, as I can now use my vpn for work on wireless again. As I’m off on a business trip next week, and all the hotels seem to be going wireless only, this comes just in the nick of time