Windows 7, McAfee Firewall and SmartPhone-Sync

December 24th, 2009 by mjoellnir

Windows 7, McAfee Firewall and SmartPhone-Sync - an Adventure. After I installed Windows7 successfully on my Samsung NC10 ("Nessie"), as described in another post in this blog, I wanted to sync my phone with it. I knew how it worked with XP. And I knew this would not help me a bit with Windows 7.

All I had, was the description how to get McAfee and ActiveSync up an running, which worked fine with Windows XP. But with Windows 7, ActiveSync is no longer required, nor is it possible to install it. Windows 7 comes with the Windows Mobile Device Center. What you probably need is the 6.1-update of the WMDC, but that's it.

At the time I was looking for information what I would need to get the SmartPhone-sync working without having to turn off the firewall completely everytime, I found nothing, no matter what keywords I fed into google. So I opened up the old description-page about McAfee and ActiveSync, and added the port that was stated there to the list of system-services:

- Service-name       : SmartPhone Sync
- TCP In/out         :   26695
- Open to a computer
   or network that is: Standard (including Trusted)

Then I looked for Apps on my computer that sounded like they might do what the ActiveSync-Apps did. I found a single 100%-match:

ceappmgr.exe,

in the directory

C:\Windows\WindowMobile .

I added it with Full Access to McAfee's Program-list. Next, I looked at the other executables in that directory. I added

  • wmdcBase.exe
  • WmdHost.exe

And tried again. It did not work. I looked through several more directories, added and removed several applications that looked promising to solve the problem, but nothing helped. Then I looked at the log-file generated by the McAfee-firewall, and noticed, that the IP 169.254.2.1 tried to access my computer through ICMP ping (which the firewall blocks), and tcp-port 990, commonly used by FTP for the control-channel when using TLS/SSL. I remembered both the IP and the port from my experiments with syncing the SmartPhone with linux, and decided to add this IP as "trusted" (which is possible directly from the log-entry).

I tried syncing again - and this time it worked.

Happy Christmas Everyone!

December 24th, 2009 by mjoellnir

Another year has (almost) passed, and it's time again to say "Happy Christmas". So, to everyone out there, have a happy, joyful christmas.

Samsung NC10 and Windows 7

December 14th, 2009 by mjoellnir

On Sunday, December 13th 2009, I dared to switch from the pre-installed Windows XP Home to Windows 7 Professional on my Samsung NC10 NetBook* ("Nessie"). Beforehand, I have had already checked whether Samsung would support Windows 7 for the NC10 or not, with the result that they actually really do. So I downloaded all Windows 7 drivers* and tools* from Samsung, made a full backup of all data on the NC10, and downloaded that little tool from Microsoft, which enables you to use an ISO-image of Windows 7, and create a bootable USB-Stick and/or DVD from it. I used my 8GB USB-Stick, and started the installation process.

The installation process went through without a problem. Even Aero started right away and the base-installation of windows was done after about 30 minutes or so - may be it was 45, I did not really stop the time. Wireless LAN (Atheros) was supported directly with native drivers, as was the sound, graphics, and most important periphal systems. I installed the drivers from Samsung nonetheless, as well as all the tools, and now the NetBook runs with Windows 7 as well as it did with Windows XP before. The memory footprint of the base-system is almost identical (approximately 600MB), and as the NC10 has been upgraded from 1GB RAM to 2GB, everything runs really nice atm.

What seems to be way faster is the SMB-stuff. When initiating a connection to a shared directory on a remote machine, I was used to some waiting before the actual content-listing could be viewed in XP's explorer window. Now the listing "is just there" as soon as I hit enter after typing the host's name - though on the remote host nothing was changed.

All in all, my first impression of Windows 7 is all-positive. I really like to use it, it looks nice with this Aero-design, and seems a lot safer with this "elevated-mode" one has to use when attempting to change system-settings. This will definitely slow virus-attacks a bit (at least until they find ways to circumvent this barrier).

I will provide more information about Windows 7 on the NC10 (as well as in general) soon.

For all interested: The NC10 has a performance rating of 2.1 (The rating you see when selecting "Properties" in the rmb-menu of the "My Computer"-icon on the desktop). I will provide the complete rating-result in a seperate posting soon.

*) Please note that links marked with an asterisk are links to Samsung Germany. As it seems, samsung.com redirects you to your local Samsung-Site depending on your IP. The tools are localized (thus german), but work just fine with any Windows 7 (I am using a us-english Windows 7 professional).

Update:

Microsoft seem to have pulled the ISO-Tool from their Websites, as the link above is not working anymore. It seems to be a consequence of an accusation that Microsoft has violated the GPL, under which the tool was released, somehow in the tool's code. If the accusations prove to be rightful, and if the tool will be downloadable again (one way or another), remains yet to be seen.

Webserver switched from lighttpd to nginx

December 5th, 2009 by mjoellnir

For some time now, I have been experiencing weird bugs with my webserver. I was running a lighttpd, version 1.4.22. Two annoying bugs occurred:

The first was more a nuisance when uploading files through a POST, which resulted in error 417 "Expectation failed" on the first attempt (while on the second it worked).

The second bug resulted in a reproducible denial-of-service, as it crashed the webserver. It occurred whenever a firefox-webbrowser attempted to connect to the server through HTTPS.

I hoped these bugs to be solved by upgrading from 1.4.22 to 1.4.23 and 1.4.24, but that did not happen. As I needed a solution to these bugs, I decided to switch from lighttpd to nginx.

The transition went surprisingly smooth. The configuration is a bit more complicated as it was with lighttpd, but easily set up and quite good explained on nginx's website, too.

Now, nginx runs all my websites, interacts with PHP through FastCGI, and problems seem to be gone.

New Server - Update (5): ROC - my personal royal PitA

December 2nd, 2009 by mjoellnir

Yeah, I guess it was a bit naive, but I tried it again. After successfully using the harddiscs in the Compaq DL360 without the ROC-Chip, at full speed and fully write-cached, I thought I re-insert the ROC-Chip and give it another try. I configured the harddiscs to run as a single RAID-1 drive...

...and yeah, you'll probably guess it already: it failed. The ROC-Chip seems to detect and disable the on-disc write-cache hardwired, as it seems. While installing the NetBSD operating system, the performance went down like hell, and I got furious.

The only solution seems to be a 64-bit PCI SmartArray-Controller-Card, which needs to be connected via a special internal SCSI-cable to the mainboard in order to bypass the onboard controller. I am currently trying to find it on the internet, but eBay was no help so far. May be I have found a reseller who can help - I have to contact them, yet. I hope they can provide such a cable, as the hardware-RAID seems to be a lost cause otherwise on that box, and I really don't want to create a new RAIDFrame RAID-1 on a system with only two harddrives total. Especially not, when the machine has the (theoretical) hardware-RAID capability as the DL360 has.

Well, we'll see how it will turn out.

Life signs :)

September 29th, 2009 by mjoellnir

Yeah, I know, I have not written anything during last months. Not really good for a blog, but I've been quite busy, since I started to study business informatics at the University for Applied Sciences Deggendorf parallel to my daily job. Prior to the real start of the first semester in october, a so-called "Summerschool" started in July, in which I participated. Two courses that would have been part of the normal semester have been taught in the summerschool, and thus the semester itself will be a bit more "relaxed" if one can say so (there will still be plenty of work left nonetheless, no doubt). These two courses were math and programming-basics. The exam for maths have been written last saturday, and as it currently seems, I passed. The exam for programming will be on friday, but I honestly doubt that I might fail there :) .

As you can see, I have really been quite busy... However, tonight, I updated the long since outdated pages regarding my computers, and added a new one: heimdall. Enjoy reading! :)

P.S.: Don't be afraid, I won't comment on every exam or course... :)

New Server - Update (4): Harddisks, SmartArray, and ROC

May 6th, 2009 by mjoellnir

In several postings in january, I wrote about a new server, which should replace the current server "unknown". I wrote about serious problems delaying that exchange, and these problems lay in the harddisk-performance.

As I wrote, the "new" server is a Compaq DL-360 (Yes, that is the complete model-name, it is the first generation of DL-360ies) . It came with two 18.2GB harddisks, and had a wonderful fast performance. I replaced both harddisks with compaq-certified 72GB UW-320 harddisks, and installed NetBSD 5.0 on it. While unpacking pkgsrc.tar.bz2, I wondered why the machine took so long...over six minutes for tarball of about 40MB. 2 minutes would have been more appropriate. At that time, I had no further for experimenting and analyzing, but on last sunday, I took up the matter again. I plugged the 70GB harddisks into Ymir, reconfiguring the Mylex DAC960PD-Controller to accept the new disk-pair as a RAID-1 drive, and ran

# tar -xjf pkgsrc.tar.bz2

It was over after 01:55 minutes, and that on a PPro-200. The DL-360 has a Pentium III 1266 and thrice as much RAM.

After this result, I examined the DL-360 again, searched the internet for information, and started a discussion on IRCnet-channel #netbsd. The advice given most often was "that sound like a write-back-cache-problem". And it was - or better, is.

What has confused me all the time was the fact that the machine claimed to have a SmartArray-Controller. Well, it has, but only a stripped-down one: It lacks a write-back-cache.  The integrated SmartArray-Controller is called "ROC" in the manual, which is short for "RAID-On-a-Chip". It is an addon-card, which I have removed now, and unless I can find a SmartArray-Controller-Card (yes, I know, they are sold on eBay), and more importantly the cable which connects the extra SmartArray-Controller with the scsi-backplane to override the ROC-Controller, I will install a NetBSD on the machine in a RAIDFrame RAID-1 again. In the end, a Software-RAID is better than no RAID at all, and RAIDFrame is really great. Unknown runs it since I have setup the machine the first time, which is several years now.

NetBSD 5.0 released!

April 30th, 2009 by mjoellnir

Yesterday, NetBSD 5.0 has been released.

As usual, you can download images via torrent and from the mirrors and read the full details in the release notes.

Here is an excerpt from the release-announcement-mail sent through the netbsd-announcements mailinglist by Soren Jacobsen:

On behalf of the NetBSD developers, I am proud to announce that
NetBSD 5.0, the thirteenth release of the NetBSD operating system,
is now available.

NetBSD 5.0 features greatly improved performance and scalability on
modern multiprocessor (SMP) and multi-core systems. Multi-threaded
applications can now efficiently make use of more than one CPU or core,
and system performance is much better under I/O and network load.

This improved performance is the result of a rewritten threading
subsystem based on a 1:1 threading model, new kernel synchronization
primitives, kernel preemption, a rewritten scheduler implementation,
real-time scheduling extensions, processor sets, and dynamic CPU sets
for thread affinity. Almost all core kernel subsystems, like virtual
memory, memory allocators, file system frameworks for major file
systems, and others were audited and overhauled to make use of highly
concurrent algorithms.

In addition to scalability and performance improvements, a significant
number of major features have been added. Some highlights are: a preview
of metadata journaling for FFS file systems (known as WAPBL, Write
Ahead Physical Block Logging), the 'jemalloc' memory allocator, the
X.Org X11 distribution instead of XFree86 on a number of ports, the
Power Management Framework, ACPI suspend/resume support on many
laptops, write support for UDF file systems, the Automated Testing
Framework, the Runnable Userspace Meta Program framework, Xen 3.3
support for both i386 and amd64, POSIX message queues and
asynchronous I/O, and many new hardware device drivers.

VCFe 10.0

April 28th, 2009 by mjoellnir

Yeah, another year has passed, and on Friday, May 1st 2009, VCFe 10.0 opens its doors in the usual place in Munich.

I will be there on the opening day, but this time just as a visitor. Time did not allow me to prepare my exhibits, so I decided to leave them at home. As last year, I will report about the 10th Vintage Computer Festival Europe and provide some photos.

Back online

April 18th, 2009 by mjoellnir

After one and a half days downtime, the system is back online, running happily on a NetBSD 5.0_RC3. With it the blogs and all the other services hosted and provided by this machine are available again of course.
I will provide a detailled report of the upgrade-process later, and the page about this machine will as well be updated later.