iSOblog

hi everyone!

i’m still there, and of course, from time to time, i need a new pc =)

this time i replaced my core2duo e6600 (2.4ghz) on a msi p965 neo-f with 2gb memory on it with my new:

• core i7 920 (2.66ghz)
• gigabyte ga-ex58-ud5 (with heatpipes, onboard power button, debug leds and everything…)
• 3x 2gb kingston ddr3 hyper-x
• and a zalman cnps-9900 led cooler

the machine now runs nicely, however (more…)

from my military service! after 21 weeks it’s finally done!

it would be too much to write about it, so i thought i’d just post my pictures, you can find them here: Pictures taken during my military service

well, 4 days late, but I still wanted to wish everyone a good start into the new year.

i wasn’t writing for some time, this is because i’m doing military service. i’m placed in Bülach, my Unit is called the “Uem/FU RS 62-1 Kp.1 Zug 3″. before xmas, i finished the 8th week, so there are still 13 left….

i was experiencing a new problem (well, it isn’t really a problem, it’s just annoying) when i opened an ssh connection to my server at OVH and i wasn’t active for let’s say about 10 minutes. when i wanted to enter a new command, the session always closed with the error:

Read from remote host xxx: Connection reset by peer
Connection to xxx closed.

sadly i couldn’t find out what exactly made this behave like this (i believe it has to do with m0n0wall, but the connection timeout there is set to 3600 seconds, so i don’t know why it happens), as it was never a problem until then. but i found an easy workaround that keeps the connection alive:

put this line into you ~/.ssh/config

this sends some data through the ssh connection every 120 seconds and avoided the connection from being closed.

i managed to achieve another certification, this time i may call myself “Citrix certified Sales Professional”. however, this certification seems to be easier to achieve and less important than the CCA (Citrix certified Administrator), which i’m going to achieve in summer 09.

the test for CCSP is solely consisting of online courses and tests which are free of charge. on the opposite, candidates have to attend instructor-led courses which are quite expensive for the CCA and the online tests are also not free.

i’m also preparing to get lpic-2 certified, however, this will take some more preparation than CCSP…

Dear Pascal Schwarz
LPI ID: LPI000115726

Congratulations on obtaining your LPIC-1 certification and joining the
world’s largest community of Linux and Open Source professionals.
LPIC-1 is globally recognized as the leading entry-level certification
in Linux skills and knowledge.

In completing your LPIC-1 certification you are also eligible to take
our LPIC-2 exams.  We encourage you to consider pursuing this higher level
certification.  For more information…

yees, finally I got it

i just updated my wordpress installation to version 2.5.1, seems like the backend (where you write posts etc) changed a lot =)

it looks quite nice and the editor seems to be even more comfortable than before.

woohoo, my server is going to be replaced soon

ovh made new dedicated server offers. the old offers are gone, here are the most important changes;

• Kimsufi (Isgenug) now with at least 1Gb ram (instead of 256mb), you can get a kimsufi for 25€/month!
• superplans have totally changed;
  • no more superplans with 250gb-hd’s
  • at least 4gb of ram!!
  • nearly the same price

so i just decided to rent a new Superplan2008R instead of my “old” Superplan2008 L – i get 3gb additional ram for 5€/month.

the plan is to copy the whole /boot and /-partitions from the old to the new server (using dd and netcat i think) and after the system is running on the new machine all things will get copied using rsync.

i’ll update here when the server is ready

well, i think many of you do already know; it’s really sad what happened.

http://www.noooxml.org/

some comments about the format by a czech blogger: http://blog.janik.cz/archives/2007/05/19/T20_32_07/

i just wanted to send a mail to phoronix, asking them about their linux benchmarking suite.

however, my mail got bounced immediately;

This is the mail system at host mail.pascal-schwarz.ch.

I’m sorry to have to inform you that your message could not
be delivered to one or more recipients. It’s attached below.

For further assistance, please send mail to postmaster.

If you do so, please include this problem report. You can
delete your own text from the attached returned message.

The mail system: host phoronix.com[209.62.40.52] said: 550-Verification failed for nottheirmail@phoronix.com
550-The mail server could not deliver mail to notmyemail@pascal-schwarz.ch. The account or domain may not exist, they may be blacklisted, or missing the proper dns entries. 550 Sender verify failed (in reply to RCPT TO command)

i was a little surprised and sent my email using another address hosted on the same server, which worked great. so it couldn’t be my mail server being on a black list.

after a little googling, i found this forum thread. the posters problem may be related to exchange (didn’t read the whole thread) but the part

Moreover, some mail server will reject email for domain whose ISP’s MX
record was pointing to an IP address instead of to a DNS name. If this is
your case, you can resolve the issue by changing the MX record to a DNS
name which then resolved to an IP address.

made it clear to me: while my other email address has an mx-entry in the form of mail.domain.invalid, the pascal-schwarz.ch mx entry was the ip of the mail server. after changing this to a dns entry pointing to the same address, the problem was gone.

as i mentioned in my last post, pascal-schwarz.ch (web- and mailserver, you’re using it right now) is now hosted inside a virtual machine on my ovh root server.

it works really nice generally, however i must also point out that this machine doesn’t generate too much load. it might be a problem to use vmware if the guest would really have a lot of IO operations to perform. but as long as it doesn’t do so, i like the simple way do administrate the virtual machine and the possibility to watch the boot operation in case there would be a problem.

i had to fix a little problem with the time-keeping of the guest. the clock inside the virtual machine was running too slow (about 0.5 seconds delay per minute which leads to a delay of around 12 minutes per day…). i did not want to install the virtual machine guest tools, as i thought that it has to be possible to fix the problem without special tools.

so i tried to sync the clock regularily using ntpd, which works usually really nice. however, to my surprise, this did nearly do no change. reading through the vmware-guide for time-keeping, i changed some options inside my (guest-)kernel which did also not lead to the wanted effect. however, it might be that those options are needed, so i’ll post them here:

• deactivate the option “tick-less kernel”
• deactivate preemption (on a server not needed anyway)
• make sure the guest doesn’t have a higher tickrate than the host (my host has 250hz while the guest has 100hz)

so i did some more tricks that vmware suggests, adding parameters to my kernel – here’s my menu.lst:

using these kernel-parameters in conjunction with ntpd, the clock now runs absolutely exact inside the virtual machine.