Older blog entries for mones (starting at number 66)

Migrating disk

I had in LJ the final story of the failed disk, so, having woken up in the mood of bloggin', it saves me a precious time :-). Lots of console output and boring stuff, you know, but here it goes:

The failing setup were two discs I synced manually from time to time, their partition table:

Disk /dev/sda: 200.0 GB, 200049647616 bytes
255 heads, 63 sectors/track, 24321 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Disk identifier: 0x0003b1cf

   Device Boot      Start         End      Blocks   Id  System
/dev/sda1   *           1         122      979933+  83  Linux
/dev/sda2             123         365     1951897+  82  Linux swap / Solaris
/dev/sda3             366        1581     9767520   83  Linux
/dev/sda4            1582       24321   182659050    5  Extended
/dev/sda5            1582        2189     4883728+  83  Linux
/dev/sda6            2190        2554     2931831   83  Linux
/dev/sda7            2555        2676      979933+  83  Linux
/dev/sda8            2677       24321   173863431   83  Linux

And mount points:
/dev/sda1               918322    445462    423864  52% /
/dev/sda3              9614148   5303228   3822544  59% /usr
/dev/sda5              4806904   3629392    933328  80% /var
/dev/sda6              2885780   1107812   1748652  39% /opt
/dev/sda7               918322      8256    861070   1% /tmp
/dev/sda8            171134396 135680768  35453628  80% /home

This setup, appart of the manual sync, had some issues to be addressed:

  • Because of packaging activities /var was always nearly full, so it had to be increased

  • Because of doubling the memory some months ago, there was less swap than current RAM size (2G), also something to fix

The hardware choice wasn't very difficult, as I tend to like Seagate, so balancing price, capacity and availability decided for a couple of ST3500418AS. These are SATA-II, while my motherboard is SATA-I only but aren't they supposed to be backwards compatible? Well, they are, but you have to setup a jumper to lower interface speed, otherwise the disc isn't even recognized by the motherboard.

Buying the discs had some more difficulties. First tried Alternate, but this time they pretend me to pay the SGAE[es] tax for media (which is around 12 euros per disc), despite I clearly explained these were system discs to be mounted in RAID (and the tax is supposed only to apply non-system drives). Phoned them even, but no way, so I finally rejected the discs and went Optize, which doesn't seem to have the supposedly legal problem Alternate has with declaring system discs. They were served on time and for less than 90 euros, so bravo for them :).

After having the bare metal, initially these options for migration were considered:

  • Buy a 2.5 disc, copy current data (a 250 Gb disc is enough), install the new system, copy back

  • Buy a hard disc enclosure for the remaining good disc, install the new system, use the enclosure to copy data back

  • Install new system in one disc (sda), copy data from current disc (sdb), replace old disc with second and setup RAID on a running system

But in the end I got it with a fourth option based on this later one: install a new system with all the RAID setup, disconnect second drive (like if the array had failed), reconnect and copy contents of old drive to new system, restore second RAID drive and add it again to the array, so it gets synced again. Nothing to buy and more fun to see how fast the MD rebuilds the array.

So finally this is the new partition table:
Disk /dev/sda: 500.1 GB, 500107862016 bytes
255 heads, 63 sectors/track, 60801 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Disk identifier: 0x000a9eb5

   Device Boot      Start         End      Blocks   Id  System
/dev/sda1   *           1         134     1076323+  fd  Linux raid autodetect
/dev/sda2             135         620     3903795   82  Linux swap / Solaris
/dev/sda3             621        2322    13671315   fd  Linux raid autodetect
/dev/sda4            2323       60801   469732567+   5  Extended
/dev/sda5            2323        3416     8787523+  fd  Linux raid autodetect
/dev/sda6            3417        3538      979933+  fd  Linux raid autodetect
/dev/sda7            3539       12170    69336508+  fd  Linux raid autodetect
/dev/sda8           12171       60801   390628476   fd  Linux raid autodetect

And mount points:
/dev/md0               1059264    263360    742096  27% /
/dev/md1              13456532   1703152  11069820  14% /usr
/dev/md2               8649480   1365644   6844464  17% /var
/dev/md3                964408     17632    897784   2% /tmp
/dev/md4              68248448    184216  68064232   1% /opt
/dev/md5             384497716 132355408 252142308  35% /home

BTW, after all I did buy the SATA enclosure for the remaining disc, so I have another 200Gb for pr0n ;-).

Syndicated 2009-10-29 06:24:51 from Ricardo Mones

-bash: /bin/su: Input/output error

That's today's message from bash. Fortunately seem's it's able to spawn it, su command and a bunch of others had not been so lucky.

The ps command is another of the lucky ones, an excerpt (hey! less also works) is highly insightful:

20292 ?        S      0:00 sh -c $SMARTD_MAILER -s 'SMART error (FailedReadSmartSelfTestLog) detected on host: busgosu' root 2>&1 << "ENDMAIL"?This email was generated by the smartd daemon running on:??   host name: busgosu?  DNS domain: [Unknown]?  NIS domain: (none)??The following warning/error was logged by the smartd daemon:??Device: /dev/sda, Read SMART Self-Test Log Failed??For details see host's SYSLOG (default: /var/log/syslog).??You can also use the smartctl utility for further investigation.?No additional email messages about this problem will be sent.?ENDMAIL?
20293 ?        S      0:00 /bin/bash -e /usr/share/smartmontools/smartd-runner -s SMART error (FailedReadSmartSelfTestLog) detected on host: busgosu root
20301 ?        S      0:00 run-parts --report --lsbsysinit --arg=/tmp/fileCJFXti --arg=-s --arg=SMART error (FailedReadSmartSelfTestLog) detected on host: busgosu --arg=root -- /etc/smartmontools/run.d
20302 ?        S      0:00 /bin/bash -e /etc/smartmontools/run.d/10mail /tmp/fileCJFXti -s SMART error (FailedReadSmartSelfTestLog) detected on host: busgosu root
20303 ?        S      0:00 /usr/bin/mail -s SMART error (FailedReadSmartSelfTestLog) detected on host: busgosu root

So /dev/sda is dying, now for real. It gave me some warnings two or three weeks ago, so last week I made a complete backup on its twin disk, using ddrescue, because dd wasn't able to do it without failing.

It's time, again, to seek for a couple of disks which last for almost five years... if possible.

Syndicated 2009-09-25 13:24:42 from Ricardo Mones


So, finally here we are, after a 6 hour travel by car I arrived to DebConf 9 venue on Thursday afternoon.

Weather is marvelously sunny (as expected), but not as hot as the figures may suggest (36ºC at arrival, something less today).

Still a lot of people to come (DebConf starts next week) but nearly about 50 are already here, and growing.

Syndicated 2009-07-18 00:21:42 from Ricardo Mones

Go virtual and die

Yesterday I got a interesting announce in my inbox: try our hosting solution for free during two months. The announce came from the nice people of Gandi, the registrar where I maintain my mones.org domain.

So far, so good. Applied and got my share in minutes. The features are not impressive, just the minimal: 256 MB RAM, 3 GB (system) + 5 GB (data) disk, 5 Mbit bandwidth and a 1/60th part of the processor, which is marketed as something between a Pentium III and Via C7 processor (not very informative).
What took most time was to wait the reverse DNS to be active because of my change of mind in the middle of the process :-). The system installed was Debian Lenny, of course, and took just minutes.

Less than 24 hours later you can see what happens, and I have not access to the server... despite I love the way Gandi does business those are not the things that inspire confidence in a hosting solution :-(.

Syndicated 2009-07-02 10:38:13 from Ricardo Mones

As times goes by

It's amazing to check and see how much has passed since last post. Not that I had nothing to tell, but maybe not in the mood to do it. Anyway there's not much excitement in my life lately, but looking back seems it isn't going too bad.

The project we were working at is already at production stage, and with only two or three phone calls to solve minor issues so far, which is not common, as I've heard. This is something I, as a the project leader, am proud of, and not being wrong with all the work-hours invested in testing and bugfixing ;-). The client seems to be happy with the results, and a second phase is planned, so more work waiting for our group. This contrasts with the landscape in other parts of the company, and the rumors floating around fed by the bad economic situation. Currently we already started other project for the our regional government, so we can't get bored at work until the end of year or so... Anyway, if luck smiles to me, I'll be doing more interesting things by the next year: yesterday submitted a grant application for review. Grants is a R+D program to provide funds for innovative ideas within the company. My idea is not so new, but it will be fun to investigate how to replace our Windows based SOE (i.e.: the image deployed in our laptops and desktops) by a Linux based one.

Time for lunch now...

Syndicated 2009-05-16 12:35:48 from Ricardo Mones

Facts and things that happen

Since last post...

  • A new year has began! (according some calendar systems, at least)

  • Obama is ruling the world! (yes, they could)

  • Debian Lenny is out! (congratulate ourselves!)

  • CVS packages built by hydra are uploaded nightly to claws-mail.org

  • Recent claws-mail packages are again uploaded to unstable (thanks to libetpan maintainer who uploaded it to sid)

  • Still working 10 hours a day...

  • But given current economic situation we have to congratulate for it... :-(

Syndicated 2009-02-28 09:26:01 from Ricardo Mones

back to vacation!

Getting older left me a nice flu as a birthday present. So nice that I had to take my first couple of sick leave days. The fever, which reached 39.1 Celsius degrees at some moment, didn't let me do other things than stay in bed and sleep (when possible). At the beginning of the week I was somewhat recovered, but today my throat still hurts slightly (probably because I've stopped taking analgesics as soon as fever disappeared).

Anyway I feel better now, and despite having missed the Claws Mail 3.7.0 released completely (no Spanish translation updates again) I can at least bring the Debian packages as a Christmas present ;-)

They're at experimental (claws-mail and extra-plugins), as previous ones, because required libetpan is still there.

This time it happened Sylpheed has released also a new version, namely 2.6.0, also in experimental now. On the other side the sylpheed-gtk1 package has been removed from Debian. I doubt someone uses it currently. Anyway, after the massive bug cleanup will let see if someone cares :-)

I have also a pending release of Clawsker, to support the new hidden preference in 3.7.0, but tomorrow we're going to visit Madrid for the weekend, so I guess it can wait until next week or even the new year! Going by car, I hope the snow and ice don't take part in the travel... wish me luck! :-)

Syndicated 2008-12-25 21:55:30 from Ricardo Mones


Hint for subject:

final class Age {
  public static void main(String args[]) {
    System.out.println (java.lang.Integer.parseInt ("14", 32));

It was last Friday, in fact, but I'm so absorbed by work lately that had no time for anything else.

Today I've been (somehow) gifted with the last album from Henry Snowstorm. In one word: captivating. In three words: craving for more. It made my afternoon despite having to waste it filling a powerpoint with useless project data for a Delivery Assurance meeting tomorrow. That doesn't mean I'm against DA, nope. But starting to do it when the project is nearly finished instead doing QA (fixing bugs) is a waste of time IMHO.

As a curiosity this post is being written from Safari under Mac OS 10.5 AKA Leopard. After being able to get a copy for free (free for me, not for its owner) I resized partitions in MacBook and reinstalled everything again (30 Gb isn't enough for Leopard upgrade from Tiger, but a Leopard installation from scratch takes much less space, which I discovered too late, after having doubled it). Now will see if I can use/build Claws Mail natively on it. For now I can see its display under the included X11 server, which seems to be a new feature in Leopard. Another one is "Spaces"... yep they have discovered multiple desktops now... amazing, isn't it ;-)

Syndicated 2008-12-15 23:50:57 from Ricardo Mones

Lenny will rock you!

It took two months, but finally the hard disk replacement for the MacBook has come. Of course I'll think twice before ordering anything else from Alternate (ES), not to mention the fact they served the SD cards without micro adapter (which means they're useless to me)...

Anyway that was not the point. The point was, as you may have guessed, I tried the latest Debian installer for Lenny, which is beta 2 at this moment:

First, the niceness of jigdo downloading the first amd64 DVD image flawlessly and without a single retry.

Second, wodim burning it without error in a disc I had forgotten for two weeks inside the drive.

Third, being able to swap broken disk and the new one despite Apple's instructions to change it forgot to mention you need a very small torx screwdriver to detach the drive from the pulling tab it has (it's located deeply inside and the tab is necessary for removing it).

Fourth, partitioning and installing MacOS X... well, last time I tried dual-boot it didn't work at all, so I had to try again, and with a 320 GB disk using less than 10% for a proprietary OS doesn't look like a great loss ;-). Partitioning is a bit tricky, as detailed in the wiki, but the Disk Utility method worked fine for me. Scheme was 30-2-288 (more or less), you'll discover later why. Worth to mention that the MacOS X showed a lot of upgrades after setting up the wireless, including an EFI firmware upgrade (new boot ROM version is MB21.00A5.B07). I installed all of these before continuing.

Fifth, installing rEFIt, though this one has no trick...

Sixth, rebooting and installing Lenny beta 2 from the DVD. To run an encrypted system two partitions are required, one for /boot (unencrypted, bootable) and other for the encrypted filesystem, hence the two partitions defined. I forgot to add a swap partition, but a swap file can be added later. The uswsusp package will warn about lacking swap, but seems it does complain even if you have a swap partition. You have to avoid installing bootloader at this point, because MBR layout is not the same that GPT (which Disk Utility wrote), hence bootloader installer would be misled.

Seventh, rebooting and entering rEFIt's disk utility, which immediately offers to resync MBR to match GPT layout. Wonderful.

Eighth, restarting Debian installation again, having to reinstall, because the filesystem inside the big encrypted partition is not recognized and had to be reformatted. Not a great problem though. Finally installing grub in the /boot partition, and finish installation.

Et voilà!

Upon restart, rEFIt menu shows both MacStuff and the penguin, and both work fine... in fact I'm writing this from the new Safari in MacOS X, because the wireless card in Debian is still to be configured, but that's another story...

Note: if you're going to try this, first of all read the wiki like I did, it has been improved a lot.

Syndicated 2008-11-12 23:53:50 from Ricardo Mones

Bits, bytes and nibbles

If you think Debian has one of the most time consuming and complicated process to join in, now look at this nice mess our beloved ftp master produced and one of the effects that caused. Despite of helping Joerg with one of his packages I'm not the one to told anybody what to do with their time, but fixing any of the RC bugs which are preventing Lenny to be released would have been more useful for Debian than writing such proposal. Introducing more classes of people in an already clustered project will only end on more problems and, IMHO, doesn't help to address the problems it's supposed to solve.

Syndicated 2008-11-06 18:51:43 from Ricardo Mones

57 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!