Justin's Journal

Uverse gone wrong

Tue, 08 May 2012 14:43:52 GMT

Yesterday morning I woke up to find that my internet service was no longer working. Go U-verse. It seems it went down around 4AM, and wasn't coming back. I called to have them reboot the box since I couldn't log into the router myself, and tried a hard reset. It wasn't working. They scheduled someone to come out that day, and eventually they arrived. After replacing my router, it still wasn't coming back. They replaced parts all over the neighborhood to try and fix things. Eventually he did a full reset, which made service come up for about 30 seconds, then it pulled an update and did not come back. After a bit of back and forth, and a lot of time, they finally got me working. This alone was annoying, it cost me several hours of my time, but it gets much worse.

For a bit of background, I have had my home network set up on 10.13.66/255 since 1998. The reasons that netblock was chosen no longer exist, and haven't in a decade, but I also didn't have much reason to change it. A few years back, I set up a DHCP server, though lots of pieces on my network were still on static addressing (again, no reason to change it at the time). This includes a lot of appliances (squeezeboxes, my aquarium) which don't have keypads for addressing. I have scripts written/used which go by address instead of name, I can tell you the IP of any box in my house as easily as I can tell you the name. It's a practice I have moved away from, but 10+ years of cruft that was working, I didn't bother changing.

The latest firmware for U-verse disables 10. networks from their "default dhcp configuration" Again, no big deal, I don't use their dhcp service since I can't set it for pxe, I use my own. I just go into manual configuration and set the router to 10.13.66.1 and I am done. Except that I can't. Their firmware now states that 10. addresses are not allowed for the router. Now, on top of my hours of downtime because the service wasn't up at all, I have to readdress my entire network. I was given no warning that this was coming, no time to prepare for it, if I wanted things working, I had to either put another router in front of that one, or readdress everything. Thanks again AT&T. I finally have everything working, but I have to ask why such a change would be made, with no notification to customers that it was coming. Why was I given a firmware in the first place that would brick my box? This wasn't the networking, as the reset put it back to 192.168.1 addressing and when it updated, it still bricked.

The last of the mess hit me this morning. I have a script that backs up lots of important things from my file server. I do not keep the backup mounted, just as an extra precaution. The script mounts, backs up, unmounts. I have been doing this for years as well. Laziness prevailed and I never bother to check that the mount was successful before I started the backup. This has worked for years, so I never thought anything of it. Ooops, forgot to readdress the nas frontend for backup, so my / filesystem was full on my mail/file server this morning instead. On the bright side, the script now checks for the mount before doing the backup, and emails me if the mount is unsuccessful.

Today I am hoping for a much less frustrating day.

Fedora 16 Cloud Test Day

Thu, 20 Oct 2011 05:26:03 GMT

Just a reminder that today is Fedora Cloud SIG Test day. Of particular note, we are testing EC2 images, Aeolus, Openstack, and HekaFS. We need your help! Please join us, lend a hand, test things out, and give your feedback. More information is available on the test page, and we will be hanging out in freenode IRC in #fedora-test-day.

Fedora 16 Beta images for EC2 are available

Wed, 12 Oct 2011 22:08:17 GMT

Just a quick note to let people know that Fedora beta images for EC2 are available, and we would love any testing you might have time for. A list of the beta AMIs as well as all other supported Fedora AMIs is maintained by the Cloud SIG here. Bookmark it as it will point to final images for Fedora 16 as well as all future supported releases. Feedback on these images is welcome:

Through IRC on Freenode: #fedora-cloud
Through the email list: cloud@lists.fedoraproject.org

Also, mark your calendars, the Cloud SIG has a test day coming up on Thursday 10/20. More info will be coming across soon!

Fedora 16 Virtualization test day tomorrow (September 15th)

Wed, 14 Sep 2011 21:18:15 GMT

Just a reminder that this Thursday (tomorrow) Sept 15th is Virt test day. Test cases and information are available on the Fedora Wiki. We will be hanging out in #fedora-test-day on freenode IRC. If you have any cycles to come help out, we would greatly appreciate it.
Should this message reach you after the 15th, your tests are still valid. Any bugs we can squash before release help out!

Fedora 15 EC2 Test Day Tomorrow

Thu, 04 Aug 2011 02:22:37 GMT

This is a little later than originally planned, but the Fedora 15
(yes, Fedora 15 – not a typo) EC2 test day will be on this Thursday
2011-08-04 [1].

Fedora 15 AMIs are available for testing and are listed on the test
day wiki page [1]. The tests are designed to ensure basic functionality
for the AMIs (MTA, httpd, yum etc.).

Since these tests require an Amazon AWS account, we are offering some
compensation (up to US$5) for the first 10 people to go through the EC2
test cases. This will be done on a first come, first served basis -
make sure that you contact rbergeron to verify that you are one of the
10 people or you may not get the credit.

Tim

PS – If you have the means to pay for the EC2 time or a free account,
please use that. We’re just trying to make sure that everyone who wants
to participate can.

[1]

https://fedoraproject.org/wiki/Test_Day:2011-08-04_Cloud_SIG_Fedora_EC2

Fedora 15 Virt Test Day Tomorrow (April 14)

Wed, 13 Apr 2011 21:40:12 GMT

Just a reminder that tomorrow is Fedora Virtualization Test Day. Test plans and more information for the event can be found on the Fedora Project Wiki. IRC for the event is on freenode in #fedora-test-day.

Please do come along and help out with the testing, 'cause we've got plenty to test! :-)

Fedora 14 Virt Test Day!

Wed, 22 Sep 2010 20:05:53 GMT

It's Fedora Virtualization Test Day

Thu, 08 Apr 2010 13:02:05 GMT

Just a reminder that today is Fedora Virtualization Test Day. Even if you don't have a Fedora 13 pre-release system to run some tests on, we provide livecd images. IRC for the event is on freenode in #fedora-test-day.

Please do come along and help out with the testing, 'cause we've got plenty to test! :-)

fudcon 2009 Toronto

Wed, 09 Dec 2009 22:52:25 GMT

I am back from Toronto and it was much warmer there than Dallas today. Fudcon was a great conference, even though the wireless infrastructure was not up to the task at the hotel or the conference site. Some points of interest:

- Attendance was great, and a massive number of talks were pitched for the barcamp.
- It was nice to get some face time with a number of people I don't normally see, and meet quite a few new people.
- There was a healthy interest in virtualization in general.
- Getting Fedora releases on EC2 is a priority, and there are a number of people excited to both use Fedora on EC2, and help us get it there.
- We have a lot of work to do for F13.
- Cynar (artichoke liqueur) is just as bad as it sounds.

All in all, the event was a success, with a great turn out and discussion. Thanks to those who put in a lot of time and effort to organize things and make sure it ran as smoothly as possible.

PCI-Express or USB sound card for Linux

Tue, 07 Apr 2009 04:43:40 GMT

Dear Lazyweb,
What options are there for a sound card with both analog and digital output, capable of unmolested 96kHz/24bit playback in either PCI-Express or USB, and that are supported with current alsa? A while back I switched my main workstation box. The new box doesn't have PCI slots, so my M-Audio had to go. While the onboard sound card is functional, it just does not compare.

Nice touch, better than a camera phone...

Tue, 11 Nov 2008 15:22:49 GMT

Sunday night I went to see The Legendary Pink Dots at the Granada theater (great show!). I don't go to too many shows there, but it is a very nice venue. Since they usually sell tickets direct, they have the email address of everyone who purchased tickets online. This morning I received an email from them with several professional photos from Sunday nights show.
One thing you see at almost all shows these days is people trying to take pictures of the bands with their camera phones. They tend to turn out about as well as you would expect from a camera phone in low light, basically a waste of electrons. It is a very nice change to have the venue send you good quality pictures from the show you attended. Sometimes the little things can go a long way. Maybe other venues will take note.

Linux Plumbers Conference

Mon, 22 Sep 2008 18:23:14 GMT

I am back from Linux Plumbers Conference in Portland last week. I have to say that I was really impressed with the conference. Generally speaking for Linux conferences, the talks are interesting and the hallways conversations are productive. With LPC the talks were really more working sessions, and they were productive in and of themselves. I think a lot came out of LPC and I certainly hope to see the conference continue. A big thanks to all of the organizers and sponsors.
As a result of the sessions and much work outside of the sessions, I think Foresight will see some significant speed up in boot time. As a side effect, certain configurations will "just work" with less hassle than was previously required. I will provide more updates once the plan begins to fall into place.

New rPath Linux 2.6.26.3 kernels available for testing.

Tue, 02 Sep 2008 03:39:36 GMT

2.6.26.3 kernels are available for testing on rPath Linux 1 and rPath Linux 2 systems. Please test and report any issues to rITS. To test these kernels with your systems for rPL 1, please run

conary update 'kernel=kernelbits.rpath.org@rpl:1 --keep-existing'

For rPL 2, please run

conary update 'kernel=kernelbits.rpath.org@rpl:2 --keep-existing'

These kernels can be tested in your groups or images by including

kernel=kernelbits.rpath.org@rpl:1 or kernel=kernelbits.rpath.org@rpl:2

Something about the Riverwalk...

Thu, 21 Aug 2008 02:39:46 GMT

I have been coming to San Antonio for as long as I can remember. Even before there was Sea World, or a reason(s) to "see Shamu", the River Walk was a nice little weekend getaway. Every trip to San Antonio has included dinner on the river at Casa Rio. Tonight we decided to stop by after the day at Sea World with the kids and it was wonderful. Beautiful night out, kids feeding the ducks (tortilla races!), eating delicious Mexican food, sipping a margarita, with the sound of bagpipes in the air... Wait, what?!? Yes, bagpipes. There is a street piper playing on the corner just above Casa Rio, he has been there for at least a year now. It somehow went well with the rest of the night. Next time you find your self in San Antonio, check him out.

New kernels for testing

Sun, 12 Aug 2007 19:55:24 GMT

New test kernels for rPL 1 based on 2.6.22.2 are available. Anyone who would like to test these kernels can do so by installing kernel=conary.rpath.com@rpl:1-devel. Please report any issues to Jira.

New kernels for testing

Wed, 01 Aug 2007 17:48:40 GMT

New test kernels for rPL 1 based on 2.6.20.15 are available. Anyone who would like to test these kernels can do so by installing kernel=conary.rpath.com@rpl:1-devel. Please report any issues to Jira.

Xen is officially in the distro...

Wed, 01 Nov 2006 14:33:39 GMT

Many of you have been following the Xen work in rtao, and wondering when Xen support would be officially in rPath Linux... Well, it is officially here! And having this inside the distro makes things much easier. Now most people cooking their groups with [xen,domU] can automatically get a xenified group ready for images, with no extra effort :) If you have your own kernel modules things get a bit more tricky, you will need to build the modules for the xen kernel as well.
Next I will be putting a 2.6.18 based xen kernel in conary.rpath.com@rpl:devel and should have libvirt and virt-manager available in rtao soon.

Test driving xen

Mon, 28 Aug 2006 20:39:35 GMT

A few people have asked about test driving xen as it stands in rtao, these are testing snapshots and the process will become much easier once 3.0.3 is released. In the meantime here are some simple instructions for testing on x86:

Starting from rPath Linux 1:
Verify that you have bridge-utils installed with conary q bridge-utils
If it is not installed, conary update bridge-utils should install it for you
next we add the xen dom0 bits:
conary update kernel=rtao.rpath.org@rpl:1[xen,dom0] conary update xen=rtao.rpath.org@rpl:1 mkdir /var/lib/xen (this should go away soon, it is a packaging bug that has been fixed but not pushed)
And if running x86:
conary update glibc=rtao.rpath.org@rpl:1[xen]

At this point you can reboot into the xen kernel and should have a running dom0. Verify by running xm list as root

Once that is working you can add your domU instances. As examples I have created a basic network server image and a basic text mode development environment. I will walk you through the setup of these images as domUs below. We have to make a couple of assumptions here, xen needs to know the path to your image file, the amount of memory you wish to give your image and the name for your domU. For these examples we will assume you wish to place your images in /srv/xen/ and you wish to give your domU instances 256MB each.
First we will set up the net-server:
cd /srv/xen/ wget http://www.rpath.com/rbuilder/downloadImage/6769/rtao-domU-netserver-0.95.img.gz gzip -d rtao-domU-netserver-0.95.img.gz

create a file /etc/xen/netserver with the following content:
name = "netserver" memory = "256" disk = [ 'file:/srv/xen/rtao-domU-netserver-0.95.img,xvda,w' ] vif = [ '' ] bootloader="/usr/bin/pygrub" on_reboot = 'restart' on_crash = 'restart'
now start your new domU nestserver with xm create -c netserver
You should see the domain booting, there is no root password on this image, so create one and use the system as you wish. To exit the virtual console, use ctrl-] To view all running xen domains on the system you can use xm list, and to reattach the console you can use xm console domID where domID is the number in the ID column from xm list

To set up the text development environmnet you would:
cd /srv/xen/ wget http://www.rpath.com/rbuilder/downloadImage/6797/rtao-0.95-x86.img.gz gzip -d rtao-0.95-x86.img.gz

create a file /etc/xen/netserver with the following content:
name = "textdevel" memory = "256" disk = [ 'file:/srv/xen/rtao-0.95-x86.img,xvda,w' ] vif = [ '' ] bootloader="/usr/bin/pygrub" on_reboot = 'restart' on_crash = 'restart'
now start your new domU development environment with xm create -c textdevel

Kernel 2.6.17 is available...

Tue, 20 Jun 2006 09:22:46 GMT

For those of you interested in 2.6.17, kernels have been cooked into kernelbits. 'conary update kernel=kernelbits.rpath.org@rpl:1' should give you what you need. As usual, if there are any problems, let me know.

New kernels...

Tue, 04 Apr 2006 06:02:14 GMT

The 2.6.16.1 kernels have been cooked into conary.rpath.com@rpl:devel for those of you interested. Lots of goodies in these and worth a spin. For the less adventerous, 2.6.15.7 has been cooked into conary.rpath.com@rpl:1. This includes a few bug fixes and one relatively minor security update. Let me know if you run across any issues with either of these kernels.

Pay no attention to that flood of commit mail :)

Wed, 22 Mar 2006 06:38:40 GMT

Over the course of the day, those of you following distro-commits have probably noticed a massive amount of email. Those would be the round 1 devel rebuild which is finally done. Group cooks and such to make it more managable will be happening tomorrow, and then devel will slow to a smaller trickle until the next major rebuild. Today's commits are around gcc-3.4.4, but include kde 3.5.1. Gcc 4.1 and gnome 2.14 will be coming soon to a repository near you :) I will also build 2.6.16 into :devel before the weekend is over.

Get your goodies here...

Fri, 16 Dec 2005 19:55:58 GMT

Lots of new goodies in the conary.rpath.com@rpl:1 repository today:
- kernel 2.6.14.4: On top of the long list of upstream updates I went ahead and restored the sata_sil/Seagate quirks for Sil 3114. Upstream disabled this for 3114 in 2.6.14 because the problem the work around was written for does not effect the 3114. Unfortunately another problem does, and this was hidden (worked around) by the code being used on the 3114. Basically this means Sil 3114 users with Seagate drives should work as well as they did on 2.6.13 and earlier.
- MySQL 5.0.16 (woohoo, MySQL 5)
- Postgresql 8.1.1
- curl and perl security fixes.
And of course the things that build against MySQL and Postgresql so that everything still works (PHP, Perl-DBD, qt)

udev/hotplug/MAKEDEV updates on :1

Thu, 08 Dec 2005 21:53:48 GMT

Currenty new versions of udev, hotplug, and MAKEDEV are being commited to conary.rpath.com@rpl:1. The update on this is just a little bit tricky because of the way udev works. Basically the /dev/ on your running system is a bit different from the /dev/ which was installed by the dev package. To properly handle this update you should run one of the following:

- If you have not yet updated udev, as root mkdir /tmp/dev; mount --move /dev /tmp/dev; conary update MAKEDEV dev; mount --move /tmp/dev /dev; rmdir /tmp/dev; conary update hotplug udev; /sbin/start_udev . This will delete all of the no longer needed dev nodes and update your hotplug and udev accordingly.

- If you have already run the update before you noticed this message, never fear, nothing is broken. To clean up all of the old unnecessary files and get things running again you can run as root mkdir /tmp/dev; mount --move /dev /tmp/dev; find /dev | grep '/dev/' | grep -v '(MAKEDEV|zero|null|console)' | xargs rm -rf; mount --move /tmp/dev /dev; rmdir /tmp/dev; /sbin/start_udev and all will be well.

Of course without running these snippits, your system should still run fine, you will just have a lot of unnecessary /dev entries. If you have missing devices you might need to run /sbin/start_udev or reboot.

fire!

Mon, 05 Dec 2005 15:05:40 GMT

This weekend I decided to introduce my 2 yr old to the time honored tradition of roasting marshmellows. Perhaps the big marshmellows are not a great introductory tool, since he took a good bite of the first one and spent the next 10 minutes getting the rest of the white sticky mess all over himself, but it was cute. He particularly enjoyed blowing out any marshmellows caught ablaze, and spearing the fresh ones for us. He seemed to enjoy the whole process.
I have always enjoyed the cold weather, as cold as it gets in TX anyway, but these days not as much as I used to. Getting dark early means I cannot take my kids to the park in the evenings when it is warm enough, and some days it is really just too cold. We had a lovely time outside on Saturday, but Sunday had just gotten cool enough that the park was not comfortable.

In rPath news, users of rpl:1 will find Openoffice.org 2.0 and Firefox 1.5 are both in the repository now. Unfortunately openoffice.org introduces a large number of new build reqs, and for x86_64 users it means group-compat32 jumps from 310MB to 900MB. It does look like the roadmap for Openoffice.org puts a 64bit clean release at 2.0.2, so things are in the works, and we can run an office suite without an extra 750MB of support libs. Of course one of the nice things about conary, is installing 32bit deps is easy, since :lib components are nearly always multilib safe by default.

rpl:1 kernel reversion

Thu, 01 Dec 2005 20:26:46 GMT

It seems that a couple more errors have popped up in the 2.6.14 kernels which were enough to bring a reversion to 2.6.13.4 up again on rpl:1. Essentially there are instances in 2.6.14 where disks will appear to be corrupt, or even worse will cause a kernel panic. This has been observed on a Sil SATA controller using an older Seagate SATA disk, which has been a troublesome combination in the past. As a result, the distro kernel on conary.rpath.com@rpl:1 has been rolled back to 2.6.13.4. For those of you who wish to continue running with 2.6.14, never fear, you can conary update kernel=conary.rpath.com@rpl:devel --keep-existing to get the latest kernel (currently 2.6.14.3)
For those of you running 2.6.14.2 from rpl:1 right now, if you start to see disk errors, you might consider a conary update kernel before you reformat, repartition, or replace the disk all together. Hopefully this will all be fixed soon.