<html>

<head></head>

<body>
    <div class="item">
        <div class="item-header">



        </div>

        <div class="item-body">

            <p>&nbsp;</p>

            <p>&nbsp;
            </p>
            <div style="height: 120px; overflow: auto; margin:20px">
                <p>&nbsp;</p>

                <p>&nbsp;</p>

                <pre>Sat Nov 27 20:57:48 exception in initAndListen std::exception: dbpath (/data/db/) does not exist, terminating</pre>

                <p>&nbsp;</p>

                <p>Sat Nov 27 20:57:48 dbexit:</p>

                <p>Sat Nov 27 20:57:48 shutdown: going to close listening sockets...</p>

                <p>Sat Nov 27 20:57:48 shutdown: going to flush oplog...</p>

                <p>Sat Nov 27 20:57:48 shutdown: going to close sockets...</p>

                <p>Sat Nov 27 20:57:48 shutdown: waiting for fs preallocator...</p>

                <p>Sat Nov 27 20:57:48 shutdown: closing all files...</p>

                <p>Sat Nov 27 20:57:48 closeAllFiles() finished</p>

                <p>Sat Nov 27 20:57:48 dbexit: really exiting now</p>

                <p>&nbsp;</p>

                <p>&nbsp;</p>
            </div>
            <p></p>

            <p>&nbsp;</p>

            <p>This, ladies and gentle geeks is what happened when I tried to start up my mongodb after an involuntary power cycle. Looking around there are plenty of posts suggesting that I should create a folder named /data/db in my hard drive. Damned if I will. Mongo db data belongs in /var/lib/mongo and in fact that's where the previously created database was placed by the system.</p>

            <p>So why is it looking for a /data/db folder now? probably because I typed in `mongodb` at the command line instead of doing `/etc/rc.d/init.d/mongodb start`. But then, I did try the init script but found that I couldn't connect to it so i tried the manual startup. Since that also failed I looked at the `/var/log/mongodb/mongodb.log` and found this:</p>

            <p>&nbsp;</p>

            <div style="height: 120px; overflow: auto; margin:20px">
                <p>&nbsp;</p>

                <p>&nbsp;</p>

                <pre>**************</pre>

                <p>&nbsp;</p>

                <p>old lock file: /var/lib/mongodb/mongod.lock. probably means unclean shutdown</p>

                <p>recommend removing file and running --repair</p>

                <p>see: http://dochub.mongodb.org/core/repair for more information</p>

                <p>*************</p>

                <p>Sat Nov 27 19:47:17 exception in initAndListen std::exception: old lock file, terminating</p>

                <p>Sat Nov 27 19:47:17 dbexit:</p>

                <p>Sat Nov 27 19:47:17 shutdown: going to close listening sockets...</p>

                <p>Sat Nov 27 19:47:17 shutdown: going to flush oplog...</p>

                <p>Sat Nov 27 19:47:17 shutdown: going to close sockets...</p>

                <p>Sat Nov 27 19:47:17 shutdown: waiting for fs preallocator...</p>

                <p>Sat Nov 27 19:47:17 shutdown: closing all files...</p>

                <p>Sat Nov 27 19:47:17 closeAllFiles() finished</p>

                <p>Sat Nov 27 19:47:17 dbexit: really exiting now</p>

                <p>&nbsp;</p>

                <p>&nbsp;</p>
            </div>

            <p>&nbsp;</p>

            <p>Pity the poor sobs who wasted their time creating directories and messing up their directory hierarchy. All your need to do is to clean up the stale lock</p>


        </div>
    </div>


</body>

</html>

dbpath (/data/db/) does not exist - but you don't have to create it

<html>

<head></head>

<body>
    <div class="item">
        <div class="item-header">



        </div>

        <div class="item-body">

            <p></p>
            <p>Having <a href="choosing-a-processor..html">chosen an AMD Phenom II as the processor</a>, it's time to choose a motherboard. There are enough brands and chipsets to make even a geek feel confused. First up, I wanted it to be something other than Asus. A toasted Asus mother board is the reason that I am having to do this rebuild in the first place. There were many AM3 boards listed on eBay (but none available locally). After a bit of research, I narrowed my choice down to a <a href="https://www.gigabyte.com/products/product-page.aspx?pid=3534#sp" id="ttl_250721430553">GIGABYTE AM3 GA-MA770T-UD3P</a> and an <a href="http://www.msi.com/index.html?func=prodcpu2&amp;prod_no=1926&amp;maincat_no=1#menu">MSI 760GM-E51</a>. I was low on caffeine at the time so I chose the MSI, now in hindsight, I am beginning to the think the Gigabyte would have been better. It was a tough choice really.</p>
            <p>First of all, choosing an AM2+ mobo instead of an AM3 board would have been the cheaper option and AM2+ boards are tried and tested while AM3 is still relatively new. Yes you can use an AM3 chip on most AM2+ boards even though the board HCL will not mention this! On the other hand AM2+ is already obsolete and the next set of AMD processors are sure to be incompatible.</p>
            <p>There are many AM3 chipsets available but getting the hottest chipset with a load of features that I will not use was not my intention. There are boards with onboard RAID 5, or dedicated memory for the GPU and multiple firewire ports. But they also happen to be expensive. The MSI board had a 'buy it now' price of $49.00 (while shipping cost was $40.00 (expedited shipping mind you)). The cheapest available GA-MA770T had bidding starting at $65.00 and shipping was $30.00 so it would probably end up being a lot higher.</p>
            <p>The MSI version comes wtih on board VGA/HDMI and an E-SATA socket, on the downside it's limited to 5 SATA sockets (I already have four drives attached to my desktop and have plenty of drives lying around). The Gigabyte board had 6 SATA sockets but no E SATA but it did have two firewire ports. My external harddrive , a Lacie 1TB has all kinds of sockets imaginable so whether the board has Firewire or SATA isn't a big deal. It would have been lovely to have a mother board that supported USB 3.0 but those were all costing $120 upwards. Better to wait a while and get a PCI USB 3.0 card. There are PCI cards with 2x USB 3 + 2 x Firewire for just under $12.00</p>
            <p>Anyway, what really decided me more than anything was the lack of caffeine and the fact that the bidding would end in a few hours on the MSI board. Hopefully it will reach Colombo around the same time that the processor gets here. Now to look for RAM. </p>
            <p></p>

        </div>
    </div>


</body>

</html>

Picking a mother board.

<html>

<head></head>

<body>
    <div class="item">
        <div class="item-header">



        </div>

        <div class="item-body">

            <p></p>
            <p>Intel's Core i7 processors, packed with upto six cores, and two threads per core, provide enormous processing power for desktops that was beyond anyone's imagination even four or five years ago. But the question is do you really need such power. The answer is hell no! Most people use a computer to access facebook, check their mail and word processing. You can do all that with an Intel Pentium IV (may be even a P3) or an AMD Sempron. Heck you can do that with a smart phone.</p>
            <p>On the other hand, some games have ridiculous minimum CPU and GPU requirements. On of the reasons for that that can be traced to relationships/partnerships between games developers and hardware vendors. <a href="https://www.engadget.com/2010/06/14/dell-hit-by-intels-antitrust-aftershocks-prepares-for00-mil/">Intel even paid off Dell</a> and others to buy their lousy but expensive chips instead of AMDs cheaper but better ones. They ended up paying <a href="https://www.engadget.com/2009/11/12/intel-shells-out-1-25-billion-to-settle-all-amd-litigation/">1.25 billion dollars</a> to AMD, at this rate AMD might decide that they don't need to make chips anymore. They will fire all their engineers and hire lawyers in their place, but I digress.</p>
            <p>So the bottom line is unless you are wasting your life playing video games, by the cheapest slowest CPU on the market and that will last you a good 3 or 4 years. The cheapest CPU would be an Athlon II and it's way ahead of an i3 and even low end i5. Honestly you wouldn't be able to tell the difference between a high end i5 or a low end i5 either so and Athlon II will serve you fine. But I must admit that and i7 is heck of a lot more powerful than Athlon II or even a Phenom II X6 black edition. But I don't need such power, on the other hand I run lots of virtual machines, and an IDE, an SQL database and a webserver, often all of them at the same time. While an Athlon II can certainly handle it, a Phenom II X4 is a better bet. And's that's what I picked up on eBay the other day. To be precise it's a AMD Phenom II X4 955 3.2 black edition.</p>
            <p>So why eBay instead of a local shop? Well most of the local shops are run by crooks and besides they don't sell AMD (in this part of the world, anti trust litigation is unheard of). The other option is to pick one up from Malaysia or Singapore but this is cheaper and I bet the quality is better too. Anyway, why am I writing this? Beause <a href="down-from-three-computers-to-two-computers..html">all three of my computers packed up at the same time</a>. Two of them are now running again (albeit on life support).</p>
            <p></p>

        </div>
    </div>


</body>

</html>

Choosing a processor.

<html>

<head></head>

<body>
    <div class="item">
        <div class="item-header">



        </div>

        <div class="item-body">

            <p></p>
            <p>Yesterday's Lakbima newspaper carries an article saying copyleft is a figment of <a href="http://saadhu.com/">daya dissanayake's</a> imagination. You can bet your hard disk that this guy laboriously wrote that article with pen and paper and then went with it to the post office to send it off to the editor. Because if he had access to a computer or phone connected to the internet, he would have first typed in 'copyleft' into the Google search box and clicked through to the first result - the <a href="https://en.wikipedia.org/wiki/Copyleft">Wikipedia page on copyleft</a> and discovered that it's not a figment of anyone's imagination, least of all my father daya dissanayake's. Incidentally Wikipedia itself happens to be a copyleft work.</p>
            <p>Recently a bunch of guys launched what they call the first Sri Lankan e-book library. All the best to them, but I have no connection with it what so ever. Neither does daya dissanayake, but he did make the keynote speech at their launch and he spoke on the subject of copyright or rather how the current copyright laws only seem to benefit publishers at the expense of authors and readers. This speech was picked up by several newspapers, and carried last week. Now one Bobby Boteju, who calls himself a writer but just happens to be a translator has chosen to shoot that down and to protect the publishers' interests. </p>
            <p>Sadly for him, he forgot to spend 30 seconds on research and made a fool of himself. Hang on a second, this guy is (or was) a Director of the Sri Lanka Performing Rights Society. Worse still, he has even written a book on copyrights! well no wonder copyright, and copyright protection is in such a sad state in this country. Kind of proves the point daya dissanayake was trying to make.</p>
            <p>Maybe I am being unfair, maybe Boteju can't afford a computer because he's spent all his money paying royalties and license fees to the publishers of the works that he translated. Being a staunch supporter of the copyright regime, that's what he would do even though it's said that many translators do not. His decision not to use a pirate copy of MS Office running on a pirate copy of MS Windows should be applauded. That's what many people in this country actually do. In fact <a href="http://www.nationmaster.com/graph/cri_sof_pir_rat-crime-software-piracy-rate">Sri Lanka ranks at number 6</a> when it comes to software piracy. Yep, a healthy respect for copyright indeed. </p>
            <p>I am sure Boteju would not have used an Indian film star on the cover of any of his books like most writers seems to. Sometimes book covers are scenes straight out of Bollywood movies, they don't even bother to Photoshop (wonder how many legal copies of Photoshop exists in this country, there are millions of pirate copies). Makes you wonder if the book is based on the Bollywood movie too. Don't be surprised if it is. </p>
            <p></p>
            <p></p>

        </div>
    </div>


</body>

</html>

Copyright or Copyleft?

<html>

<head></head>

<body>
    <div class="item">
        <div class="item-header">



        </div>

        <div class="item-body">

            <p>&nbsp;</p>

            <p>Just a few weeks back I mentioned that <a href="geek-porn-unpacking-a-macbook-pro..html">I had three computers</a>. It never pays to boast because now I am down to two. For a very brief period, I was actually down to just one. All three of them had failed at different times and two of them went together. But two of them have been restored. The travails caused by one of these is the topic of a <a href="from-p5b-se-to-p5b-vm-se.html">recent post</a>. That box was my media/file server. It died. It's been sick for a long time but finally threw in the towel just over a month ago. Cause of death; a faulty motherboard.</p>

            <p>There after it was the turn of the second computer, my desktop. Fortunately it wasn't anything serious just a hard drive failure. Hard disk failures are an <a href="000840.html">annual event</a>. I am so used to it that I didn't loose any data at all. I simply removed one of the drives from the defunct media server and place it into the desktop. In fact Palimpset had been reporting at least one drive on both those computers to be faulty but I choose to ignore it for quite a while (except for religious backing up). So it wasn't one drive that had to be chucked out but two. Incidentally, both drives were the newest drives I had. The older ones are fine. They just don't make them like they used to.</p>

            <p>&nbsp;</p>

            <figure style="text-align:center"><a href="http://images.raditha.com.s3.amazonaws.com/blog/images/hw/p5g-mx.jpg"><img alt="Faulty motherboard" height="267" src="http://images.raditha.com.s3.amazonaws.com/blog/images/hw/p5g-mx-t.jpg" title="p5b-vm SE" width="400"></a>

                <figcaption>Replacement board - as bad as the original</figcaption>
            </figure>

            <p>&nbsp;</p>

            <p>Finally it was the turn of the Macbook to act up. Macs are nothing but trouble, why people pay so much money for cheap Chinese crap I don't know.</p>

            <p>&nbsp;</p>


        </div>
    </div>


</body>

</html>

Down from three computers to two computers.

<html>

<head></head>

<body>
    <div class="item">
        <div class="item-header">



        </div>

        <div class="item-body">

            <p>On <a href="asus-p5b-se-toast.html">October fourth</a>, nearly a month ago, I tweeted about a toasted Asus P5B SE mother board and about breaking my life long habit of not claiming a warranty. It turned out to be such a waste of time. The owner of the shop from which I bought the mobo, is a decent bloke. Wish I could say the same about his staff. When I sent the motherboard there, the guy at the counter had refused to accept it saying the warranty had already expired. it had not, I spoke to him over phone and persuaded him to accept it. That's par for the course in Sri Lanka. About two weeks later he called me up to inform me that I would be getting a replacement - well that certainly is unusual, the norm is that you have to call them they never call you.</p>

            <p>Finally got the replacement about a week ago, the P5B SE is no longer being manufactured so they had sent a P5B VM SE. I think the blame for the delay probably should goto the local agent (Epsi) rather than the shop. Things went smoothly at first, until the machine hibernated; it wouldn't resume afterwards. This is not the first time that I've had trouble with hibernate so I naturally put this down to an OS/software issue and disabled power management in both the BIOS and the OS. Again things worked fine for a while, then once Chrome died and took X Server with it. The architecture of the X system would make that nearly impossible but it did happen.</p>

            <div style="height: 200px; overflow: auto; margin-top: 20px; margin-bottom: 20px;">
                <pre>Oct 31 07:35:06 localhost kernel: abrt-hook-ccpp[2797]: segfault at 0 ip 00007faef8a8b022 sp 00007fffdd9fd098 error 4 in libc-2.12.1.so[7faef8a0b000+175000]
Oct 31 07:35:06 localhost kernel: Process 2797(abrt-hook-ccpp) has RLIMIT_CORE set to 1
Oct 31 07:35:06 localhost kernel: Aborting core
Oct 31 07:35:06 localhost kdm[1697]: X server for display :0 terminated unexpectedly
Oct 31 07:35:07 localhost kernel: NVRM: Xid (0001:00): 53, CMDre 00000000 00000000 00000000 00000001 00000001
Oct 31 07:35:08 localhost abrt[2800]: saved core dump of pid 2469 (/opt/google/chrome/chrome) to /var/spool/abrt/ccpp-1288490706-2469.new/coredump (261390336 bytes)
Oct 31 07:35:08 localhost abrtd: Directory 'ccpp-1288490706-2469' creation detected
Oct 31 07:35:09 localhost abrtd: Package 'google-chrome-beta' isn't signed with proper key
Oct 31 07:35:09 localhost abrtd: Corrupted or bad crash /var/spool/abrt/ccpp-1288490706-2469 (res:5), deleting
Oct 31 07:35:11 localhost kernel: irq 16: nobody cared (try booting with the "irqpoll" option)
Oct 31 07:35:11 localhost kernel: Pid: 0, comm: swapper Tainted: P           2.6.34.7-61.fc13.x86_64 #1
Oct 31 07:35:11 localhost kernel: Call Trace:
Oct 31 07:35:11 localhost kernel:   [] __report_bad_irq+0x3d/0x8c
Oct 31 07:35:11 localhost kernel: [] note_interrupt+0x118/0x17d
Oct 31 07:35:11 localhost kernel: [] handle_fasteoi_irq+0xa8/0xce
Oct 31 07:35:11 localhost kernel: [] handle_irq+0x88/0x91
Oct 31 07:35:11 localhost kernel: [] do_IRQ+0x5c/0xc3
Oct 31 07:35:11 localhost kernel: [] ret_from_intr+0x0/0x11
Oct 31 07:35:11 localhost kernel:   [] ? mwait_idle+0x7a/0x87
Oct 31 07:35:11 localhost kernel: [] ? mwait_idle+0x2c/0x87
Oct 31 07:35:11 localhost kernel: [] cpu_idle+0xaa/0xe4
Oct 31 07:35:11 localhost kernel: [] rest_init+0x7e/0x80
Oct 31 07:35:11 localhost kernel: [] start_kernel+0x413/0x41e
Oct 31 07:35:11 localhost kernel: [] x86_64_start_reservations+0xb3/0xb7
Oct 31 07:35:11 localhost kernel: [] x86_64_start_kernel+0xf8/0x107
Oct 31 07:35:11 localhost kernel: handlers:
Oct 31 07:35:11 localhost kernel: [] (nv_kern_isr+0x0/0x5e [nvidia])
Oct 31 07:35:11 localhost kernel: Disabling IRQ #16
Oct 31 07:35:36 localhost kdm[1697]: X server startup timeout, terminating
Oct 31 07:35:06 localhost kernel: abrt-hook-ccpp[2797]: segfault at 0 ip 00007faef8a8b022 sp 00007fffdd9fd098 error 4 in libc-2.12.1.so[7faef8a0b000+175000]Oct 31 07:35:06 localhost kernel: Process 2797(abrt-hook-ccpp) has RLIMIT_CORE set to 1Oct 31 07:35:06 localhost kernel: Aborting coreOct 31 07:35:06 localhost kdm[1697]: X server for display :0 terminated unexpectedlyOct 31 07:35:07 localhost kernel: NVRM: Xid (0001:00): 53, CMDre 00000000 00000000 00000000 00000001 00000001Oct 31 07:35:08 localhost abrt[2800]: saved core dump of pid 2469 (/opt/google/chrome/chrome) to /var/spool/abrt/ccpp-1288490706-2469.new/coredump (261390336 bytes)Oct 31 07:35:08 localhost abrtd: Directory 'ccpp-1288490706-2469' creation detectedOct 31 07:35:09 localhost abrtd: Package 'google-chrome-beta' isn't signed with proper keyOct 31 07:35:09 localhost abrtd: Corrupted or bad crash /var/spool/abrt/ccpp-1288490706-2469 (res:5), deletingOct 31 07:35:11 localhost kernel: irq 16: nobody cared (try booting with the "irqpoll" option)Oct 31 07:35:11 localhost kernel: Pid: 0, comm: swapper Tainted: P           2.6.34.7-61.fc13.x86_64 #1Oct 31 07:35:11 localhost kernel: Call Trace:Oct 31 07:35:11 localhost kernel:   [] __report_bad_irq+0x3d/0x8cOct 31 07:35:11 localhost kernel: [] note_interrupt+0x118/0x17dOct 31 07:35:11 localhost kernel: [] handle_fasteoi_irq+0xa8/0xceOct 31 07:35:11 localhost kernel: [] handle_irq+0x88/0x91Oct 31 07:35:11 localhost kernel: [] do_IRQ+0x5c/0xc3Oct 31 07:35:11 localhost kernel: [] ret_from_intr+0x0/0x11Oct 31 07:35:11 localhost kernel:   [] ? mwait_idle+0x7a/0x87Oct 31 07:35:11 localhost kernel: [] ? mwait_idle+0x2c/0x87Oct 31 07:35:11 localhost kernel: [] cpu_idle+0xaa/0xe4Oct 31 07:35:11 localhost kernel: [] rest_init+0x7e/0x80Oct 31 07:35:11 localhost kernel: [] start_kernel+0x413/0x41eOct 31 07:35:11 localhost kernel: [] x86_64_start_reservations+0xb3/0xb7Oct 31 07:35:11 localhost kernel: [] x86_64_start_kernel+0xf8/0x107Oct 31 07:35:11 localhost kernel: handlers:Oct 31 07:35:11 localhost kernel: [] (nv_kern_isr+0x0/0x5e [nvidia])Oct 31 07:35:11 localhost kernel: Disabling IRQ #16Oct 31 07:35:36 localhost kdm[1697]: X server startup timeout, terminating</pre>
            </div>

            <p>On closer inspection it seems that it's not just X server that is dying. It could be caused by faulty hardware, on the other hand it could be OS related as well. I've seen <a href="2228.html">gnome and GDM freeze freeze up</a> like that while using Fedora 10 on another box. So I switched to KDE with no luck, then I tried LXDE still the problem persists. Namely the machine either gets stuck or the screen just goes blank - the only way to get it back is to power cycle. So the only conclusion is that the replacement motherboard is also faulty. Well it didn't come with the original factory packaging, I'll bet it's something that they recieved as part of a trade in deal. And I am pretty sure they are not going to replace it again.</p>


        </div>
    </div>


</body>

</html>

Monthly Archives for 2010 11

dbpath (/data/db/) does not exist - but you don't have to create it

Picking a mother board.

Choosing a processor.

Copyright or Copyleft?

Down from three computers to two computers.

From P5B SE to P5B VM SE