Older blog entries for berend (starting at number 366)

15 Jul 2011 (updated 15 Jul 2011 at 02:08 UTC) »

Unfortunately cmdbulkimp destroyed the KnowledgeTree database, existing documents were moved all over the place. So what now?

Maybe I'm a moron. It appears there is an upload from server link in the admin interface.

Too bad Google didn't return this as the top result.

13 Jul 2011 (updated 15 Jul 2011 at 02:04 UTC) »

Have been doing some work with the KnowledgeTree document management system. Was about 61GB of data, so got the company to sent a disk to Amazon to import. As I didn't know you could import to EBS (or maybe they just announced that) this was imported to S3.

It's a pain to get your data out of S3, let me tell you that, especially as in their case they had deeply structured files. s3fs can mount this, but didn't see any files. In the end I used the Perl Amazon::S3 library to get all files in the bucket and download them one by one.

Although the data is now on the server, it doesn't mean that it's in KnowledgeTree. There is a Best Practice document, but the the link to the bulkimporter is dead. Was able to locate it. Now seeing if this is going to work.

If you ever get:

svn: Commit failed (details follow):
svn: Commit item '/...../some_dir' has copy flag but no copyfrom URL

Try this:

svn delete ..../some_dir
svn revert -R ...../some_dir

Or else the slightly less safe:

rm -rf ..../some_dir
svn update ...../some_dir

Just completed some new kind of work for a Manukau motor lodge, new for me that is: SEO improvements. Got the help from the New Zealand search masters.

Was running on Ubuntu 9.10 for a long time, and very happy with that. But one must move on I suppose, so had a bit of downtime yesterday and decided to upgrade to 10.10. Ubuntu upgrades always went fantastic, never had a problem, but this time it was a nightmare. Had upgraded a couple machines already, but not yet my main machine.

First upgraded to 10.04, the LTS release. Everything went OK, until I booted. When it came up, the nvidia driver didn't load, so into low res graphics mode. Weird. Booted on an older kernel, that went OK. As I had to upgrade to 10.10 anyway, I didn't bother to pursue it. And somehow I got the idea that I should do the upgrade from the latest kernel, so I booted back into the latest kernel with low res graphics. But I think the problem was that this kernel didn't support my Lenovo W700 well, so the kernel simply didn't work and that's why nvidia didn't load.

I could start the upgrade to 10.10 fine, no issues there. But about half way X window/machine/kernel crashed. Could no longer boot.

On Windows this would have been a reinstall, but luckily on Ubuntu you have options. So I downloaded the Ubuntu 10.10 live CD, booted from that, and then chrooted to my disk. With liberal use of "apt-get upgrade", "dpkg --configure -a", etc. and several retries I somehow was able to finish the upgrade and install the 10.10 kernel, which worked fine.

System still in a somewhat bad state though. mysql wasn't installed, so had to do that myself. Bit of a problem as well as my my.cnf had statements in it that prevented it to start. Starting it gave me all kinds of warnings that I needed to upgrade.

Lost my Thunderbird profiles, currently loading them from a backup tape.

And then of course we have the usual and expected: PHP upgraded to 5.3 which will cause issues. I'm first trying to see if I can resolve them, before going back to 5.2. For example SimpleTest doesn't work anymore, but the latest trunk fixes this.

25 Sep 2010 (updated 25 Sep 2010 at 07:13 UTC) »

Recently I experiencing a 2s to 3s lag with Skype calls. Finally found out why. This is what you see when you have this lag:

$ pacmd list-sink-inputs | egrep '(current|requested)'
current latency: 2004.02 ms
requested latency: 66.00 ms

The solution is to remove loading the module-suspend-on- idle in /etc/pulse/default.pa.

12 Sep 2010 (updated 12 Sep 2010 at 23:23 UTC) »

Had a really weird issue with:

sieveshell localhost

core dump would follow. So I tried to debug this by putting print statements to see what function would fail. That brought me deep into the bowels of Cyrus IMAP and next into Cyrus SASL.

In the end the final function that failed was gss_init_sec_context() which appeared to be a FreeBSD function. And that finally gave me a Google search on the the problem. It's a very long thread, but the solution is here.

Another would be to disable gss.

Not sure if Jeremy Chadwick's patch is already accepted, because, if not, with the next cvsup the problem will come back.

Michael Brandon is trying to get links for quadracentifiable. Good luck with that Michael!

Source upgrade from FreeBSD 6.4-STABLE to FreeBSD 8.1-STABLE completed. Only one issue though:

ad2: 152627MB <WDC WD1600BB-00GUA0 08.02D08> at
ata1-master UDMA100
GEOM: ad2: geometry does not match label (81h,63s !=

This appears to be a harmless message and can supposedly be fixed with:

bsdlabel -e -A /dev/ad2

And it did. After manually fixing the heads, I could mount it. Ran an fsck:

fsck /dev/ad2e

to make sure nothing was screwed, and it wasn't.

Just finished reading a very beautiful article in CACM July 1010: x86-TSO: A Rigorous and Usable Programmer's Model for x86 Multiprocessors.

Very useful, very clear semantics.

357 older entries...

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!