Faster file copying over a local network

No Comments

What I started to do was to copy a 460MB file from the computer in the project network into my laptop before I leave in an hour. The problem is that when I started mapping the drive containing the file and copying it using Windows explorer the ETC displayed by the file copy dialog box fluctuates between 2-5 hours. This instability and slowness is a pet peeve I have against the Windows copy process over CIFS/SMB shares. I know copying from the command line is going to be faster but without a progress meter I don’t know when it is going to be finished.

The next option I did was to share it the webserver hosted in the remote box and then tried to download it. The download speed range is between 30-75kbps so the ETC reported by the browser is between 2-3 hours. It may be faster but not fast enough for my needs. Then I realized that I have rsync inside the cygwin install in the laptop. As I have already the network share mapped to Z: drive so the simple command

$ rsync -avp --partial --progress /cygdrive/z/Temp/bigfile.zip ./

started the transfer which was a whopping speed range of 230-420Kbps. ETC is now 15-20 minutes. Problem solved so I am writing this blog while waiting for the transfer to complete. For those interested with the rsync options:

  • -avp : copy the file in archive mode, preserve permissions and increase verbosity.
  • –partial : if the error fails, don’t delete the transferred chunks so it can be used for resuming the transfer.
  • –progress : show the copy progress.

Why didn’t I just use a USB stick? It is because the company has locked down the USB storage drivers and while somebody who amazingly has a strong resemblance to me can do something about it, it would be too troublesome to pursue. FTP might have been a good choice of protocol as this is a secured environment but from experience the antiquated file transfer protocol only has a slight advantage over HTTP in terms transfer speeds.

ciao!

Semi-hosed Wind UMPC (Fixed)

No Comments

Fixed by overriding the GRUB lines, editing the GRUB menu, and correcting the entries in the /etc/fstab.

Here is the original line that causes the bootup to panice:

kernel (hd1,0)/boot/vmlinuz BOOT_IMAGE=linux root=/dev/sdc10 vga=788
initrd (hd1,0)/boot/initrd.img

PCLinuxOS is configured to use the graphical GRUB menu so pressing the ESC key will allow the users to go into text mode that allows the editing of the line command. Pressing ‘e’ on the selected entry enabled to correct the order of hard drive being searched [ (hd0,0)/boot] as well as the correct root [root=dev/sda10]. After both the kernel and initrd lines have been updated, pressing ‘b’ will continue the boot process.

The GRUB entry is only half of the problem since the boot operationg still panics and dumps me in a command line. The culprit is that the fstab file still contains incorrect entries since all of the partitions are being searched at /dev/sdc. The USB flashdrive I used for the installation has 3 partitions so there are also 3 incorrect entries for it in /etc/fstab.

As I said, the boot process dumps me into a command line. This is where knowledge of vi comes very handy. It is small enough and ubiquitous enough to be in most *nix system. A quick ‘vi /etc/fstab’ to correct all linux partitions (and temporarily disable mounting of windows partitions) and I am now good to go. Invoking ‘mount -a’ confirms that I am able to successfully mount all partitions.

This problem is solved but unfortunately the PCLinuxOS Minime distribution is not the answer to my wireless in the Wind woes. 🙁

ciao!

Semi-hosed Wind UMPC

No Comments

Doc Mana (or rather his daughters) own a Wind and has pointed me to the released Linux realtek drivers. As Doc Mana has stated, the released drivers are in RPM format but this gives me a good opportunity to use the alien utility for converting packages. Doc also said his Neo empriva is working fine with a similar wireless chipset.

With the extended weekend, I promised to try it out with a refreshed Xubuntu installation, or better yet try installing the PCLinuxOS MiniMe distro to get native support for RPM. The nutshell is that I now have a semi-hosed Wind UMPC. 🙁

Here is what happened:

  • I replaced the gnome network manager with the kwlan application and I hosed down my network connections. I then decided to refresh the Xubuntu installation from the my 4GB flash disk. And then…
  • I tried converting the package using the alien utility with the “–scripts” parameter because it told me so. It said it is for including the scripts inside the package. When it was time to install them with dpkg, an error was thrown out that there is a parsing error in the post install script. Both drivers show this error. Great!
  • As this is a refreshed installation, I decided to try again the tip from the msiwind.net wiki as this is also stated in the Gentoo Wind wiki. The build essential tools refused to install properly because of the problem with the botched realtek drivers installation.
  • I cannot uninstall/purge the realtek drivers from synaptic becase there are errors in the hooked scripts. 🙁
  • I wiped by flash disk with MiniMe and proceeded to install over my Xubuntu system. Everything seems to be fine until I rebooted. The system got reconfigured with my flash disk as its main storage (sda) and my Wind’s drive as its secondary (sdc) so the GRUB entries refuses to work.
  • I tried booting with the flash disk inserted and tried to select the HDD Minime installation but the boot process refuses to proceed and ends with a kernel panic.

Right now the Wind will only continue booting if I have the flash disk inserted, and its too late into the night to start tinkering with livecd options. The bright side is that I have effectively created a very crude security key. /lol

I hope I can fix this properly tomorrow as the wife is giving me weird looks when she can’t easily get into the Windows partition to play Mystery Case Files. 😀

ciao!

Ext2/Ext3 undelete

3 Comments

Frequently I come across queries on how to retrieve files in Linux wherein the most probably answer will be that it is almost impossible due to the journaling nature of the standard file systems for most Linux systems. Ext2 is not a journaled file system but its use for large partitions (>128MB) is not recommended since a premature crash (system-induced or external like a power shut-off) will cause the file system to be marked as dirty and be subjected to a file system check and possible data loss. At this time where terabytes of disk space are becoming available to us members of the unwashed masses, I don’t think anybody in a right mindset will even consider using ext2 on a 20GB partition.

Back to the topic, previously you are screwed if you were clicker/presser happy with delete operations (rm -rf ~ /tmp comes into mind, ouch!) then it only takes a split second from the time you confirmed the delete operation to switch between nonchalance and an agitated mental anguish. It may be a single document or your whole directory (I hope you are not in the habit of running as root for day-to-day use), but if you really need to get that file you just delete and you are using ext2/ext3 then there might still be some hope. I just came across the page for ext3undel which promises a set of utilities and scripts that will help you in getting your precious files back. The project is not available on official Mandriva and Arch repositories 🙁 but the project provides RPM and Deb downloads.

I haven’t used it yet as I am fortunate enough to haven’t gotten the need but the documentation on the project’s wiki seems to be simple enough. I am going to try installing this on my systems this weekend since this would be a handy tool to have when my stupidity arises need beckons.

ciao!

tux on a stick

No Comments

I have a 4gb TDK usb drive and i want to try running a full-blown linux on it. I am using this guide from pendrivelinux to install xubuntu on the flash drive.

The instruction is for creating two partitions: the first one is a fat16 partition that will hold the main system while the second partition is a small ext2 partition for the persistent data (settings?). I modified it a bit with this setup:

  • sdb1 – 3GB FAT32 – for sharing data
  • sdb2 – 1GB – fat16 for the main system. I tried using ext2 but the instruction was really for a fat16 partition for the syslinux step
  • sdb3 – 115MB – persistent partition that is labeled casper-rw

The partitions were created using cfdisk on my archlinux system. I opted to put the 3GB partition at the start because I am from the old school thinking that when dealing with stupid OSes always put the partitions it can read first. All of the partitions are physical partitions but i marked the second one as the active/bootable partition.

Everything else in the tutorial went smoothly except when it was time to boot the machine using the usb drive. The booting stopped with the message that the boot sector was invalid or damaged. I tried googling around for the fix which basically suggests installing the boot loader on the flash drive’s boot sector (/dev/sdb) but these did not work.

I finally gave in and restarted the installation process, only this time I placed the 3GB partition at the end so that I will be mostly compliant with the pendrive linux guide. There is really something to be said for following instructions especially if those instructions do not contain any discussion on the rationale. My tux in a stick works although it takes quite a while to boot my laptop. I have been able to use it in the office so I can copy a big file (RTC linux client) that I need to bring home from the office since I cannot complete the download at home.

I am happy since I have a portable Linux system in case I need it. As for the question on whether Windows was able to recognize the data space at the end of the drive, I honestly don’t know. I haven’t booted into Windows yet here at home, and the USB ports are locked by the domain server at work. I guess I will eventually know the answer but I am not in a hurry to find out. 🙂

ciao!

Fixing the swap partition

1 Comment

After a few failed attempts in getting my ArchLinux desktop system go into hibernation, my boot-up shows an error during the mounting of the swap file. Issuing the “free” command validates that I don’t have any swap space running. My desktop system has 2GB of physical RAM but I wouldn’t want to wait for something to fail before fixing the swap space.

The fix is to format the swap partition (/dev/sda8) again before activating it using this command sequence:

# mkswap /dev/sda8
# swapon -a

mount -a would also work in place of the swapon command.

NOTE: mkswap will happily format the device or file you pass on as parameter so make sure you are passing the correct partition or you will be sorry. 🙂

ciao!

Third time is a charm: Mdv Spring 2008.1 is in full swing.

2 Comments

Since the start of this week I have been struggling with the installation of the latest Mandriva release which I have just finished downloading via torrent. I even posted a passing comment about it in the unofficial Mandriva Users board. Since I have already gone through two succeeding installations that failed, I stated that if the third time didnt push through then I would really consider looking at another distro for Lala (my laptop).

The first failed installation has something to do with a recurring error in the wifi prism54 loading that made it impossible to login. I don’t have the exact error with me but it has something to do with the eth0 not being in a final state, and the IRQ channels might be too busy. The second attempt wherein I tried to only install the bare minimum of X package borked the whole X installation. I installed Mdv again yesterday and this time just selected the KDE installation. If this one is also going to fail then I wouldn’t waste my time manually choosing the packages to be installed. I left the installation running while we had our dinner. When I came back the installation was already showing the root account and user configuration window.

At the final configuration screen, I opted to just configure the ethernet connection and setting it to start at boot but not at that specific point. I just want Lala to boot up and then I would configure the wireless card using ndiswrapper since I already have the Fujitsu provided drivers in a CDRW. Taking the installation DVD out of the drive bay, I crossed my fingers for the first reboot.

For the remainder of the boot sequence, I was pretty much holding my breath until I got to the KDM screen. I entered my user account and waited for KDE to kick in. The good news is that everything seems to be ok but it was taking quite long for the KDE startup sequence to complete. I think I waited around 5 minutes which is not good. I noticed that KDE has started by default a lot of daemons that starts in the taskbar How very MS-like so the first task was to remove them one by one until only the clock and the network applet remained in the status bar area. I don’t really understand why I did that since I plan on replacing KDE with XFCE4, my new favorite desktop that replaced fluxbox.

I put in the CDRW disc and copied wireless card Windows driver in /tmp. Weird thing is that ark is not able to extract the file so I ended up doing a manual unzip operation in the command line. This is not a good start but I can live with it. I proceeded to configure the wireless card using the network applet. The package manager asked for the DVD installation media since it needs to fetch the ndiswrapper package so I swapped in the DVD+RW in the drive and let the urpmi front-end do its thing. After a minute I was asked to locate the Windows driver’s INF file so I navigated to the /tmp directory. After a few default choices, I was shown the Wifi access points in my area which listed only my wireless router ( I live in a pretty low-tech neighborhood. 😀 ). After a few test, I was greeted with a message that my network has been properly configured.

Hurray! But still I have no connection. A quick ifconfig command later and I saw that I have a wlan0 device but it doesnt have an IP. I just issued a quick dhclient wlan0 and voila: a wlan0 device with an attached IP address! I took advantage of the situation and downloaded the latest packages in the Internet as well as Krusader and XFCE4. Everything finished at around 12AM so I made sure everything was finished and halted the machine.

Tonight I booted up Lala and pressed the escape key to see the bootup messages in the background. In the network start phase, there was an error in the “Set Encode” portion of the wifi loading so I was pretty disappointed. I logged in using the XFCE4 session and was inside my desktop in less than a minute after I entered my username and password. My biggest prize was waiting inside: the top XFCE4 panel has the network applet and it is showing 3 out of 4 bars. My laptop is connected automatically to my wireless router! SUCCESS!

I am now writing this post using my spanking new Mdv installation. I am still amazed that Mandriva is consistent in getting the middle trackpad buttons working while I am struggling to do it in Windows. I just need a NAS server and everything should be set for the file coordination between Lala and Leo (my desktop Archlinux installation).

I am happy even if I am already sleepy due to lack of enough rest. Anybody who wants a copy of the latest Mandrake DVD (Free Edition), let me know but you will have to meet me in Robinson’s Pioneer during office days. 🙂

ciao!

Growing Old and Growing Up

2 Comments

The Philippine Linux User Group (PLUG) mailing list currently has a thread that (according to the archives.free.net.ph) is 44 replies strong. The thread started about a query on how to increase one’s technical portfolio (I assume for job interviews, but it is still too early to trust my memory) but, in typical internet forum style, morphed into opinions and lamentations that the PLUG lists are getting boring and losing the old hands.

Various comments were put forward but the reasons that stuck with me were that (1) the usability state of Linux has evolved that solutions to majority of technical issues are just a Google away, and (2) the old posters have moved on to the pressing need of their real life.

I don’t see myself much of a contributor but I have been a member of the list for more than five years and I am not posting as much as did. I don’t even police the lists anymore but I have always attributed that to the reduction of the number of PLUG mailing lists, the members sharing the policing responsibility, and the members actually knowing better than to post off-topic questions. The last two signifies that the PLUG mailing list members have grown up and the new members are savvy enough to respect proper list netiquette or they are afraid to post, whichever you want to believe in. 😀

People might not grow up proportionally to the rate they are growing old but they will grow up nonetheless even if they don’t realize it. The PLUG thread sparked an eureka moment, I am growing up. I just noticed that I now often sign my blog comments using my nickname instead of just “ramfree17”, and if left unchecked I am using proper capitalization on those comments which is a far cry from the “all low-caps” rule that I was imposing when I was younger 😉 . The downside if it can be called that is that I am losing some of my idealism in favor of pragmatism. I was always in favor of pragmatism since childhood but I guess right now the pragmatic part of me is consuming more of the idealistic share.

An email reply jokingly contributed that the “golden years” of PLUG might have already passed but it is time for the new blood to step up. Whenever I hear the term “golden years” I remember this snippet from Mary Schmich’s essay “Advice, like youth, probably just wasted on the young” (made popular as Baz Luhrman’s “Everybody’s Free (To Wear Sunscreen)”):


Accept certain inalienable truths: Prices will rise. Politicians will philander. You, too, will get old. And when you do, you'll fantasize that when you were young, prices were reasonable, politicians were noble and children respected their elders.

ciao!

A night of interoperability

4 Comments

Last night was my first night of putting interoperability of Linux to a personal test. I work for a Microsoft-centric company so every time I need to do active meeting sessions with counterparts then I have to boot into Windows to ensure that nothing goes off in the middle of the meeting. This is because meetings like those entail a lot of opening of project office documents, searching mails for references, and other stuff. The project documents are also a work of their own since I think the person creating the base templates are certified MS Office template designers (if there is such a thing).

My new project requires me to handle late Tuesday night calls since we are conferencing three very distributed time zones (Manila, Madrid, California). Conference calls are done using Skype since I don’t have IDD capability at home nor am I willing to do the call on my prepaid SIM. Since I already have Skype installed yesterday on Archer, I tested a conference call using Skype’s echo123 service. I was able to hear my recorded message so that portion is not a problem. My headset has a small static sound but I think it has something to do with the connection between the headset and the speaker audio out port. The only possible stumbling block would be the mail searching portion and the documents that needs to be referenced in the call.

An hour before I have already downloaded all reference materials off the web mail interface of the company’s Exchange server. That reminds me that I have the Exchange server because it keeps on kicking my Firefox connection whenever I open a message. The sidebar is OK but the main area gets a “Connection is lost” message. IE doesn’t have this problem when I am in Windows but I wouldn’t install ie4linux just for this now that Metrobank already supports non-IE browser. If anybody knows how to make Thunderbird connect to an Exchange server then let me know although the configuration might also be painful. I couldn’t even configure Outlook in my desktop to connect to the Exchange server *even* after using the company supplied configuration utility. I might even try using Evolution during the weekend. 🙂

Anyway, I just created a new Firefox Windows and opened all possible mails that may be referenced in the call so I wouldn’t search for it. I then tried opening all Excel, Word and PowerPoint attachments using OpenOffice.org (2.4). OOo was able to open them all which is admirable enough but I have the following nitpicks:

  • For MS Word document some of the formatting especially on tables are misaligned.
  • The OO.o Writer is showing too many artifacts like ruler margins and such that could be done without.
  • Default rendering size for the spreadsheets were too tiny.
  • The comments on the Excel spreadsheet are showing far from the cell they are attached to. I have to scroll just to see them.
  • Default rendering size for the presentations are also tiny. I had to play with the zoom control for a bit until I finally hit the “Optimal” option. Why wouldn’t the Optimal option be selected as the default?

Understandably, the nitpicks above are minor and there might be some preferences that could be set to remedy them but I am not using OO.o that frequently aside from a view-only app. The resulting documents from OO.o saves require too much effort to correct to make it my mainstream application for work documentation.

Back on the topic of the call, everything went relatively smoothly although my Skype cannot connect nor message the contact from Madrid. I can see his status is online but for some reason our Skypes just couldn’t communicate. The conference call had to be initiated by the contact from California.

This whole blog post might be seen as petty since there isn’t any groundbreaking achievement but I don’t care as this is a personal achievement about Linux and Windows interoperability. While Linux cannot fully replace Windows right now, Windows’ share of the home computing time is getting smaller.

ciao!

Philippine Linux User Group Zeitgeist

No Comments

I just learned that German word and wanted to use it in a post. Upon Google search reveals that it meant current intellectual trend. 🙂

The current Philippine Linux User Group site is experiencing a slow renaissance of some sorts thanks to the efforts of Gabriel Mercado and company. The latest addition to the site is the aggregation F/LOSS related posts of some PLUGger bloggers. In a post to the list, Gabriel posted the FAQ of the aggregation as


- *What the heck is this?* - This is the aggregated feed of a few
gentlemen PLUGgers / Bloggers who've been kind enough to share their Linux /
FOSS or otherwise PLUG related feeds with the rest of us. Namely, they are
Messrs. Anuerin Diaz, Elijah Alcantara, Manny Amador (Phnix), Michael Cole,
Migs Paraz, Noel Bautista and Zak B. Elep.

- *What's it for?* - Several purposes. Primarily, to achieve 'zeitgest'
so that we may be able to know what things are going on in and around
PLUGgers minds, what's keeping them occupied, and so on. Secondarily, it
provides content for our website. Once syndicated, said content will also
help promote PLUG by spreading our news, info, howtos, etc., onto the
general blogosphere. It can also help us figure out who are actively doing
FOSS work out there, as well as provide a central repository of all things
PLUG related. Not to mention of course, end hunger and promote world peace.

Take note however that each post still belongs to each author and does not reflect the official stand of PLUG (the organization). If I mess up here in the capsule then the blame will all stop here.

Subscribe now! It’s a fun way to get an insight on what is currently happening on the local Linux mindsphere. Unless you have a local RSS reader, I am recommending the Google Reader since it is free and pretty nifty.

ciao!

Older Entries Newer Entries