My 10.04LTS servers are nearing the end of the line, but the move to get them upgraded is starting to get close....
At first I was waiting for the 12.04.1LTS update to appear, so that I might try upgrading them to Precise. But, then it struck me that it struck me that it might be time to replace one of the servers with new hardware.
The boxes had been servers for many years now, but in 10.04 they dropped support for 32-bit hardware as servers....which means one of my servers, and the more important one...is nearing EOL.
So, back in August....I came across a Shuttle XS36V open box on newegg.com....so I ordered one, along with an 8GB SODIMM memory set (a pair of 4GB)....the computers support a maximum of 4GB, but in my mind I was thinking that if this worked, I might get another to make a pair.
But, then it sat around as I go on to other projects....I had figured on using one of the 120GB SSDs that I had acquired earlier....
Realized what was strange about these boxes was they had serial ports, so I started thinking about network serial and setting these boxes up headless. Since I had done a server at work for serial with FreeBSD, I got to thinking that I would do the same with this. Plus I was really finding to be much more predictable/stable than Ubuntu....something that's good in a server.
I started revising the plans as I went....ended up deciding that I would have one that is default cable and one that is default dsl, and running the usual core...one would be master DNS, and the other one be the primary slave....just as one would be primary DHCP and the other secondary....both would do freeradius. Both would NTP servers. Both would do mail, though zen is my main mail system. They would have both apache and nginx on them...apache to run local web services....and nginx doing reverse proxy to the other web apps on my home network (kind of important since current cable router is a step back with it only supporting 16 port forwards....though I have a replacement that I'll get to setting up some day...)
Then while I was poking around on newegg.com, spotted that there was another open box Shuttle XS36V....so now I had two boxes. I had picked up once Cyclades ACS1 earlier on ebay, so I got a second....a 'new' one for about the same price. Just needed to find time. Perhaps after FreeBSD 9.1 drops.
Well, it dropped....but I still wasn't ready...now I needed some SSD drives for the machines. So, I used the $30 in RewardZone for the new TV I had gotten in September, and picked up a pair for 128GB SanDisk Extremes. And, then I started setting up the ACS1s....hmmm, wonder if I have any serial cables. Okay, I'll buy some of those first.... I end up ordering from TigerDirect, as they have a better price on RJ45 to DB9 adapters....not that I need them for the actual ACS1 to Shuttle XS36v connection...that's just a 9F to 9F null modem cable.
I was poking around on the Radioactive@home message boards, after finding that my sensor had stopped working when the app was updated from v1.69 to v1.72....it was complaining every 11 seconds with "No heartbeat from client for 30 sec - exiting".
And, it wasn't because I had to build my own newer boinc client for box last month. The Ubuntu apt version is 6.12.33, and didn't seem to be any moving to upgrade it for 10.04LTS or in any PPA that I could find. And, downloading the binaries from the boinc.berkeley.edu site didn't work, as they depend on newer libraries than exist for 10.04LTS. So, I had to grab the source
git clone git://boinc.berkeley.edu/boinc.git
and build that against my 10.04LTS system....and that worked (at least the client did, which is all I really needed)...until now.
Though I read in forums that v1.72 wasn't supposed to be released (very little testing had been done), but it was needed to support newer sensors which are out in the field now.....oh well.
Well, I supposed the HEAD of the boinc might be too bleeding edge as well. 7.0.44 is considered the development version, and not recommended. While the recommended version is 7.0.28.
The wiki page on getting source says:
Various versions of the code are available at any given point:
The master branch contains the latest source code for all components of BOINC. It is always under development, and has not necessarily been thoroughly tested.
The client software (client and manager) is maintained as follows:
The code for each release is 'tagged'. For example, the code for version 6.3.14 is tagged with client_release_6_3_14.
A 'branch' is created for each released minor version. For example, the branch client_release_6_2 is the code for the latest release of version 6.2.
Except I tried to
git checkout client_release_7_0_28, and it couldn't be found. Nor could I do
git checkout client_release_7_0. Eventually, poking around on the site some more, I found that the tags were different. They uses '.' instead of '_' in version separators. Though couldn't get to 7.0.28 directly...had to get 7.0 first.
git checkout client_release_7.0 followed by
git checkout client_release_7.0.28. Got me the source for the current recommended version.
Now at 7.0.28, I couldn't get
_autosetup to complete...missing
py/Boinc/version.py.in. Well, copy those from the HEAD down, seemed to kind of work...though configure was having a lot of issues. Lots of SAH_LIBEXT and SAH_DLLEXT command not founds, along with others. Turns out files were missing from the m4 directory. Compare with HEAD, and copy over
Makefile.in, ax_lang_compiler_ms.m4, sah_links.m4, sah_largefile_breaks_cxx.m4, sah_libext.m4, sah_select_bitness.m4, and sah_staticize_ldflags.m4.
Now I do configure, this time I look to see what things its complaining about missing, and add more -dev libraries, along with wxWidgets....now I have a boincmgr, should I choose to use it.... Though the compile didn't go well.
Missing a couple of header files....
client/dhrystone.h. And, then I go and blunder with doing a
make install, which clobbers my
/etc/default/boinc-client with something that doesn't fit the Ubuntu environment. So, once again a quick restore from BackupPC of these two files, and BOINC is working again.
While writing my post on the Radioactive@Home forum, I spotted a new chart.... Which I have now made into the footer of this site:
However, what was interesting...was going back through the last couple of weeks, to find this span....
Maybe the world was supposed to end on December 21st?
That is Gumby was torturing me!
It all started around May 13, 2010....and could ended around...sometime back in April, 2011...but had to let it drag on to today...July 23, 2011.
Gumby had been torturing me more and more lately....largely because when I upgraded it in May 13, 2010...it became my second fastest system at home, and my fastest Windows XP box. Ole TARDIS was only a Pentium 4 3.2GHz w/HT. While new Gumby is a Pentium D 3.4GHz.
Though while TARDIS was still my main Windows XP system, I largely kept Gumby's role the same...though it did pick up a few things the ole Gumby didn't do, due to it having more juice and being less cluttered in its registry and driver hell. Like a new DVD burner, or running a few more CPU and/or memory intensive apps (things that didn't work to well on P3-933MHz and/or only a max of 512MB...)
But, the problem with new Gumby was that its a legacy free box. No PS/2 mouse or keyboard ports (though it's strange that I can still buy new systems that have one or two PS/2 ports, something about the reliability of PS/2 mouse and/or keyboard over USB probably... this KVM switch was kind of holding up whether or not I would be requiring PS/2 ports on whatever mobo I decide to build the new LHAVEN around.)
I had gotten USB PS/2 adapters to keep using my old keyboard/mouse and KVM setup on Gumby (though the keyboard and mouse have upgraded a few times since the Trendnet TK-408) Also, I think the mouse has always been a USB kind with a PS/2 adapter. Originally a ball type, now an optical type....some day I'll get some laser mice here and there...
Problem was that the USB PS/2 adapter was less than reliable or usable. I was still using the Kingwin one that I got to start with new Gumby. Which worked the best, except that every now and the the mouse would stop responding. I could have to stop using the mouse for a while, and it had to be absolutely still during the stop....and then something would reset and it would work again. I had tried a TrendNet, thinking it would go with the KVM....but switching away and then back on the KVM would cause the scrollwheel functionality to go away, and that required me to unplug/replug the USB adapter to get scrollwheel working again.
I also tried a PCI PS/2 card, but found that internally was just another USB PS/2 adapter...just in a slot consuming form, plus it didn't get along with the other PCI card (firewire, and I use firewire external storage on Gumby)...also it exhibited the same problem that the TrendNet USB PS/2 adapter did, except that its a little harder to unplug the card to reset it at will...
So, I stayed with the KingWin and set my sight on a KVM.
This is an owed post (of over 3 months in the making....) I had thought of lots of things to write for this since I meant to write it, but didn't write anything. And, now its so old...that I'm just going to be really terse....
The lead in to this posting was Worked on Orac yesterday. So, I got those 5 2TB drives...and first tried to make it RAID5 using mdadm. And, started filling it up....though the performance didn't seem to be much better using XFS, plus the new backuppc seemed to have regressed on handling XFS.... so I heard about ZFS on Linux....as kernel modules that do the taint. And, tried out RAIDZ, with its variable striping to maybe to do better on performance. It seemed nice, but then stack dumps filling up my logs....rebooted and then did a zpool scrub, and it went to work fixing lots and lots of errors, and then a few unrecoverable errors.
Guess no ZFS, and the XFS didn't seem enough of a gain to make it different from everything else. Even though there's one really big problem with ext4. Resizing it causes corruption. Supposed to be fixed now, but before I switched to the new array, I grew the backuppc volume for 3.75TB to 3.8TB, offline...and it resulted in corruption still.
So, I was making the new backuppc volume the full size of the array. As I mulled over the situtation...I decided that I would try RAID10, sure that wouldn't get as huge a boost over the old RAID6 array of 1.5TB, but it would still be an increase...and no idea what I'm going to do when I need more space. But, first I had to get a 6th 2TB drive to do RAID10. Plus it would mean that I would need to do something else with the 1.5TB drives. One was going to have to come out. Sure it could run degraded initially (though did find that it would come back after a reboot, guess the boot making degraded arrays come online...only does it for OS filesystems. Though the boot get's stuck because it can't mount filesystems and the recovery of mounting just doesn't work still. Getting into single user is such a pain too. Though after I broke my sudoers file recently, I turned off hiddenmenu.
The hard part about making the RAID10, was figuring out the ordering of devs...so that it was mirror across the eSATA channels and stripping with in. There's only 2 channels, and 6 drives in this array...so figured that was the best way to go. Though now that I've realized that its only PCI express 1.0, not sure if that was the best way to go. Though to try the other, would call for me to get 8 drives? And, who knows what the future holds for supporting backuppc....
internal bitmap on doesn't seem to degrade the RAID10 as much as RAID5 or RAID6, so I have that turned on now.
Now there was the question of what to do with the 4 remaining drive bays and the 4 1.5TB drives. Would I got with another RAID10, a couple of RAID1...would I concat them into the LVM for growth or make them separate.
Well, on the old 1TB array was my old MMC volume, which needed more space and part of the 5x1.5TB RAID6 was a backup volume...also in need of more space. I decided that I would go with two separate RAID1s using the Seagate 1.5TB drives (I had pulled the Samsung 1.5TB to make it degraded....that drive has since been hooked up to TARDIS for local backups).
Around this time, my Roku appeared...and I went through the various attempts to map a network drive to it for local content. Settling on the HSTI Media Stick and its 1TB maximum. I settled on going with 2 1.5TB RAID1 arrays. One of them was made fully available to be the new MMC volume. And, the other...I made a 1TB volume (called TARDIS) for the HSTI Media Stick, and the rest became the new backup volume.
The old 1TB RAID1 array stayed online, until it was recently repurposed into the volume for Time Machine backups (only 931GB).
Now I've been debating playing around with external bitmaps and/or external journaling to see if I can get more performance. But, external bitmaps go away after a reboot, and that just seems to be a bug that isn't going to get fixed anytime soon. external journaling would require me to find some devices for that....so I got the idea of PCI Compact Flash adapter and some compact flash cards....I have a couple of the PCI Compact Flash adapters....and I do have a few compact flash cards around. Though when I was playing around with ReadyBoost on Zen, I found they weren't really that great on speed. And, I was going to want to get as much speed out of things as I could here....so I did some checking and now I'm waiting to get around to buying some 600x CF cards to get around to trying this. The fastest card I have available in my collection is a 133x. And, I don't recall why I bought that one. Ones earlier than that would be left overs from when I had my PowerShot S20, then an REB1200 and then a Nikon Coolpix 5700.
These days, I'm all about the SD cards for my digital cameras (Eye-Fi), though been thinking of getting an ultracompact (and wearing my holster less often)....and some that I've looked at use MicroSD cards. Don't have any extra Class 10 ones laying about though....though I do have a couple of Class 6 8GB cards that need a home (I got them in a 2for1 sale, and the need for 1 didn't happen as planned....and when I do get around to getting that device, I'm sure I'll probably cave and get a 16GB or 32GB Class 10 MicroSD card.)
So, the other day, I was poking around on NewEgg.com, trying to decide if I wanted to get the upcoming ShellShocker. I spotted a single slot GT440 for a good price among the daily deals. While I was looking around, I saw other things that caught my interest...a powersupply tester. Or things that I needed to go with the card, like a DVI-I to VGA adapter. Also decided to grab a new fan, since it is probably only a matter of time that the fan in the other drive array on Orac to fail.
I ended up not getting the ShellShocker item.
I've been eyeballing getting a single slot 400 Series card for Orac for a while. Was originally going to get a 200 Series, and I did actually get one in a different ShellShocker deal. But, when I opened up Orac, I realized it wouldn't fit....due to its huge fan.
Not sure what I'll do with that card. But, I had been looking at 400 Series, because Zen came with a GT420....which I've learned is a pretty horrible, compared to the rest of the line. Especially when it comes to doing BOINC. I'll probably go GTS450 or GTX460....single slot, though I need to take a look inside someday to see if there's actually juice in there. Otherwise, maybe I'll want a second one of these...
So, I had picked up the card from the UPS Store on Friday...but I didn't get around to installing it today. I stopped checkarray before I shutdown. I changed the scheduling of checkarray to spread things out better.
Though I had a brief distraction this morning on another matter, which I'll be posting about soon...
But, I went and put it in...its a pretty tight fit sitting next to my PM eSATA card.
I wonder if I'll look into a slot fan to go in the open PCI slot there....the lower PCI slot is where 4 other eSATA....oh wait, I had gotten a PCI card that I was going to put into Orac. I wonder what happened to that project....
Pages: 1· 2
In the aftermath of the summer storm of August 13th, (hmmm, totally missed that it was a Friday the 13th), I made a tweak to my ddclient config for updating dyndns for my DSL line. Because I found that it wasn't able to update the IP change while Cox was down.
Couldn't find a way to make ddclient to bind to the local IP that routes out by DSL (or use non-default gateway). But, since I have squid proxy on the same box...and depending on what port I come in on, it can use either of my connections.
I set proxy=box.lhaven.homeip.net:3128
Couldn't use localhost, because ddclient does some kind of validation to require an fqdn+port, and localhost isn't an fqdn. And, yes, I use my dyndns domain as my home domain. So I can have bookmarks that'll work whether I'm at home or on the road
But, this change wasn't tested...as it has been less than 28 days for a refresh, and no IP change.
That was until this morning, when my IP did change.
The updates weren't working....seems that ddclient wants to do SSL all the way or not at all. No using an http proxy to connect out on SSL. But, I didn't feel like sending my dyndns password out non-SSL.... So, after some thought, I decided I would figure out how to set up SSL on squid.
I made the necessary configuration change, but no go. Seems that ubuntu doesn't distribute squid with SSL, because squid and openssl have incompatible open source licenses. So, I did a quick search to find the ubuntu way of rebuilding it from source.
apt-get source squid apt-get build-dep squid apt-get install devscripts build-essential fakeroot cd squid-2.7.STABLE7 vi debian/rules Add --enable-ssl \ to “# Configure the package” section debuild -us -uc -b cd .. dpkg -i squid??? squid-common???
Change to proxy=box.lhaven.homeip.net:3218, and it worked.
AKA the end of RedHat 7.3.
For sometime, I have thought that it was time to upgrade 'lhaven' to a more modern release of Linux.
For the longest time, 'lhaven' has been running RedHat 7.3 (it was 7.2, but then I discovered Fedora Legacy...and they were providing patches for RedHat 7.3) And, there are still some production servers running this release, but I no longer have to support them...as I no longer work there.
The problem was that I had compiled a lot of custom apps over the years, that would be difficult to upgrade to newer distro/release of Linux. Since they were done without though to going to a different distro. 'lhaven' did used to run older versions of RedHat, and had been upgraded to where it stopped. But, to go to another distro would require a major effort. And, some of the apps were hard enough to get working to begin with. And, were some what fragile.
'lhaven' has also gone through some hardware iterations. The first major was when the Cyrix 233 that it was one burned up (a transformer had blown a phase, causing really low voltage before it went out....) though more likely it was its return after the outage with none of its fan able to spin. It was working for a while...but then there was smoke, and it all ended.
Got a Duron 800Mhz machine, which grew to an Athlon 2000....and had its motherboard replaced after the icepocalypse. The icepocalpyse was the first time I had to seriously look at upgrading from the old RedHat 7.3....but later I found a compatible motherboard and got the system to boot. The machine I bought at that time became 'orac'.
Later my use of ReplayTVs dropped off, and the one of the main apps that used datadirect for guide data was never successfully updated to support the newer schedules direct service. And, that was one of the things I didn't want to touch. But, now it is gone.
More recently, I started moving some of the other things that ran on 'lhaven' to 'box'. Namely, MRTG and cacti. I also updated all my linux servers to be able to handle mail for my home network, with the local mailboxes on 'box'. Also moved secondary DNS to 'orac' (someday I plan to get fallback DHCP on 'orac', but that'll probably come later.....) During the icepocalypse I had moved DHCP and primary NTP and DNS to 'box'....'lhaven' came back as secondary, because the old 'gumby' was sufficiently broken that I couldn't get it to not require 'lhaven' to function.
So, what I did was install ubuntu-10.04-alternate-i386 to become the new 'lhaven'. Which was the suggested to do RAID at the start, it had been a long time since I had done it, but the last time I had...it was a server install. But ubuntu no longer supports 32-bit hardware as servers, but both 'lhaven' and 'box' are 32-bit machines.
Took some figuring out, but its all done now. Yes, there's a desktop on all my Linux boxes, though 'lhaven' doesn't seem to be really up to the task. The video seems pretty poor and fuzzy compared to the other Linux servers. Perhaps I'll see about replacing its integrated video with an NVidia card of some sort...what that does CUDA, though it'll have to be a PCI card.
Anyways....after the install was done, I installed a few basics...like boinc and called it a night.
'orac' is back, long live 'orac'.
'orac' had gone away on Thursday, it actually started in Wednesday.... For some unknown reason, I decided to start the upgrade of 'orac' from 8.04LTS to 10.04LTS.
The upgrade seemed to go okay, but when I rebooted....it wouldn't boot. I scrambled around looking for a livecd to boot with....did have any 64bit ones handy and 10.04 was hit and miss on booting. But, I could see vg0 (root, home, swap) with all the livecds except 10.04.
Something about 10.04 didn't like how vg0 was set up. I looked at 'box', which had been running 10.04 for a while and was working. Noticed the main difference between the all the other md's and vg's...was that they weren't fdisk'd md's. When I had originally set up 'orac', I had used a recipe that I had found online on how to do it...since initial install of ubuntu onto Raid wasn't straight forward then. and after md1 was created, I had fdisk'd and then pvcreate/vgcreate, etc. on it.
When I added the other two raid sets....I just pvcreated the md and then used it...no fdisk of the md.
In retrospect I should probably have done a backup of the filesystems while I was running a livecd that could see it....and then perhaps remade the disk and restored into it. But, I didn't and while trying to 'fix' the problem...I corrupted things, and I wasn't able to recover (tried pvcreate -u + vgcfgrestore...but couldn't get the filesystems to come back.
After lots of other attempts and such, I reached the point where I decided things were lost and set on to figure out if I can get back.
backuppc localhost doesn't back up the entire localhost....but it had my home directory backed up (but not all of /home), and it was also backing up /etc. Realized the the one thing it wouldn't get is my crontab files. But I had a dd image of most of the disk (the front part munged)....so I string'd and recovered my crontab from it. And, then set out to install 10.04LTS new.
Since md1 was going to get redone, I decided to redo md0..../boot. Made it a bit bigger than before so that it won't get as crowded from accumulating every new kernel that is released until 12.04LTS comes out. Or some future release needed more than the old size to work. Had actually run into this problem years ago at my previous job....
But, the install kept failing....asking for the disk again....I burned several copies, download from other sources, but no go. So, I tried to 9.10 (thinking I would upgrade to 10.04 immediately), but it also didn't work. Eventually, I found a reference on google that said the 10.04 (and 9.10) installs were picky about what kind of optical drive was being used. It specifically mentioned a particular DVD burner, with 'orac' came with a DVD burner...though I have never used the drive since I did the original install. I tried a USB stick, but couldn't get the machine to use that, and the other external drive I could find was firewire (and there's no BIOS option to boot from firewire).
So, I took apart my machine and the firewire drive. And, put the DVD-ROM drive that was in the firewire enclosure into the machine. The DVD-ROM drive had original come from my circa 2003 Dell, known as 'tardis'.....when I had swapped in a DVD burner.
'box' is the Linux machine that I do cd burning on....and I hardly do DVDs anymore, though I did get a new DVD burner recently for the new 'gumby'.
Anyways....with the 'new' optical drive in 'orac'....I was able to get 10.04LTS to install, and then started the long process of getting my environment back. First problem was backuppc installed with a different uid/gid....so it wouldn't access the old /var/lib/backuppc storage. I opted to fix it the wrong way....that took forever....doing the chown of the files/directories. Should've just changed the uid/gid that backuppc had installed itself as. But, then I tried to do the restore....next problem was sudoers, had forgotten about that.
I then restored /etc to another place, and restored my home directory along side of the one the install had created. When done I swapped my home dir and rebooted. That was a mess. The desktop was all messed up...not just missing apps. Figured it was because the desktop didn't like the old settings....but couldn't find any details on how to upgrade desktop, if I hadn't done an upgrade. I thought other OSs handled this kind of thing. But, not here. So, I flipped back and slowly copied stuff over and restored settings etc. And, reinstalled various apps as I went.
Then I decided that I had made sufficient progress on getting 'orac' configured, that I didn't want to lose things. I had disabled backupc from doing automatic backups during the restore. So I turned it back on and back up the other systems around my home (after first doing a new backup of the new 'orac'). It didn't like .gvfs in my home directory....in old version it complained but continued, in this version it complained and the tar returned an error due to previous error...causing the rest of the backup to fail (the localhost backup does 4 'shares')....so I tweaked the config to get it to work.
The only thing I saw was that it tried to backup 'ulkc', my ubuntu notebook...but it wasn't on at the moment, so it failed. It wasn't going to automatically back up anything else yet, since its one of my blackout periods. So, I started kicking off manual backups of various things, and watched it work. First thing that didn't work was a backup of a windows share. Seems smbclient changed since 8.04LTS...and one of the options was used in backuppc 3.0 didn't work anymore. The new is 3.1, and the default options is minus that one switch....so I matched things up and it did its thing.
The only problem that still didn't work, which I think was the reason I decided to bite the bullet and upgrade to 10.04LTS (to get the newer backuppc) was incremental backup where a very large file has changed since the full. Usually the virtual harddrive of a VirtualBox. Sadly this wasn't fixed in 3.1. Guess I'll have to look at some other way to deal with it. Like exclude it from the regular policy and create a new one that only does that one file....or something. Or maybe ....
I continued to work on other things, as I worked through comparing the old /etc with the new 'orac' /etc. Though interrupted by trying to get file shares working (I had exported my Dropbox directory from another server for access from other computers on my home network....though LAN sync is available now, I haven't undone the stuff yet). And, getting snmpd working again....so I could see the impact to resources and network from all the backuppc jobs running in cacti.
One strange problem is that 'orac' can not access the NFS shares on 'lhaven' (RedHat 7.3), but 'box' can....the only difference I can think of is that 'box' was upgraded to 10.04LTS.... Though I guess it is time I worked on moving services off of 'lhaven' and figure out what I'm going to do to either replace or upgrade it.
The other issue was Handbrake. Originally, I had built it by hand...since there wasn't a package for 64-bit Linux. Though later, I got a newer version via PPA. But, the version that I got via PPA for Lucid didn't work. So, I eventually went and built my own again...from the subversion....my build works....
Hopefully, when 12.04LTS comes along....things'll just upgrade smoothly.
The other odd thing, is that 10.04LTS seems slower....particular in file I/O. The machine seems to really degrade in performance doing a large file copy (as was happening when I was copying directories from my old home dir to the new one.) All the CPUs are bored while this is happening (it's a 2.4GHz Core 2 Quad....with 8G of memory....) Hopefully something that'll resolve itself later?
Last night I wasn't thinking....running into the fact that my boinc-client package is broken on 'box', I thought..."why isn't it running the boinc-client from Lucid?" And, I proceeded to fix it with "apt-get"...and then realized that I'm still running karmic and unbreaking it wasn't. The boinc-client pre-Lucid is too old for a project I'm in, and it is annoying how ubuntu doesn't update packages within releases (which get's really annoying with LTS.)
Anyways....I decided the way to resolve the mess I had now made by downgrading my boinc-client, was to upgrade to Lucid Lynx 10.04LTS. It was my plan to upgrade both my Ubuntu servers to 10.04LTS (the other being 'orac' which is currently 8.04LTS)....but I was going to put it off to when I had more time and allow time for the release to stabilize (given all the issues I had when I first started running box on fresh Karmic).
So, it started....things got a little annoying, in that the upgrade requires attention....and I didn't want to spend my whole evening watching it upgrade. So, I'd check it now and then, and sometimes find it stuck waiting for me to make a decision for it, though the first one where it wants to restart stuff after a pam upgrade...that's annoying. Just do it okay, stuff it restarted weren't working right along the way anyways. Namely I found that I couldn't access websites...at first I thought the squid had become dorked, and I couldn't restart it because invoking /etc/init.d/squid said it was now a service...but using service didn't find it. Though later I realized it wasn't working because DNS was broke ('box' is my primary DNS server)... I was able to restart that in the usual manner.
Partly because of the pauses, and a bit due to slow download....it wasn't looking like I would finish before 'bedtime'....but it seemed that it would be close enough.... It wasn't, but I ended up staying up to the bitter end. So, that I could reboot and do a quick check that all was clean.
Namely, I checked that named, dhcpd, ntpd and squid were running (since these were some key services of this server, and were ones that often failed to start after boot in karmic...especially named. Manual start always worked...at one time I was restarting these services in rc.local where it always worked rather than at the 'normal' time).
Then I went to fill in the missing icons in my launcher panel....there were two holes, there also seemed to be a hole in the tray area. The missing icons were Evolution (which I right-clicked to make reappear), and boxee (which I installed the latest of). I then called it a night.
The next morning, I continued to poke around some more....adjust appearance, add some chat/broadcast accounts...and look into my ubuntuone issue....first the tray thing was missing, apparently its by design...but, I don't remember that. I poked around some more and some more....then I looked at 'ulkc' and saw that it also didn't have the tray applet anymore, and I just didn't miss it... So, I then went to check that it was connected to my account and sync'd. Actually, it really was syncing, etc. When I was at Penguicon, I had put a copy of the pictures I had taken into my cloud...expecting to see them later on 'box' when I got home. But, it would never appear. Upgrading 'ulkc' didn't help things.
During the poking around, I disconnected 'box' and reconnected...and that apparently fixed whatever was wrong, as it started syncing...and soon the folders appeared.
While this was going on, I thought that maybe I would need to finish things up by remote....so sadly I discovered that remote desktop still doesn't work in Lucid. Turning it on, consumes 50% of my CPU. My 8.04LTS (orac) doesn't suffer from this problem. But, I want to bring both servers in sync eventually, so I can see about getting failover for dhcp working between the two.
Oh well....at least it seems more successful than when I had upgraded 'box' to Karmic. And, now I'm at an LTS where it can stay for the next 2 years.... We'll see what happens when I upgrade from 8.04LTS to 10.04LTS next month....
Since I got back from Penguicon 2010, it had been my plan to test out upgrading to Lucid Lynx.
I didn't get around to it the first day back to work, since I was busy catching up on 1000s of emails and making sure things that my co-workers had ignored or overlooked during my absence got done. I then came home with a fever, and ended up taking the next day off from work. The fever broke Thursday night, so I went back to work on Friday...even though head is still all stuffed up and I had a killer sinus headache.
Finally, near the end of the day, I remembered that I wanted to try the upgrade....
At work, I have two Karmic Koala VMs, running in VirtualBox on a Sun Ultra 20 running Solaris 10/x64.
One VM is/was a 32-bit server....it runs BOINC all the time, got a squid instance when I was testing something and more recently I set up backuppc so that I'd have backups of my windows desktop at work. The other VM is a 32-bit desktop which is mainly so I can offload and run my mail/calendar client on...since the Windows computer is starting to suffer with the bloat and growth of apps...the more recent versions of Thunderbird and Firefox are very resource intensive, and I can't run both at the same time anymore.
Though what I do now is run Thunderbird and direct it to open on Cygwin/X running on the Windows box. Since, I want it to be visible all the time as I'm doing tasks come from emails. Makes it strange opening links and attachments though, since they won't go to the Firefox/apps on my windows box....etc.
I first upgraded the server VM, and it largely worked without any major issues....the only thing is that it does complain with "Your CPU is lacking expected security protection"....the checker says the CPU family has NX, but my BIOS isn't ... but the box to enable PAE/NX in VirtualBox is checked. So, guess its just a bug with Ubuntu running in VirtualBox.
I then upgraded the desktop VM...again that went okay.
The server VM originally started out as a Jaunty Jackalope, and went to Karmic Koala...so that was my first experience that 32-bit servers are no longer supported by Ubuntu (I mentioned it as the Ubuntu Q+A at Penguicon, but was ignored...when they were saying that Ubuntu supports a wide range of older hardware..I have older hardware that won't run 64-bit that I use a servers at home...plus I ran into a lot of problems getting 32-bit apps to work on my 64-bit hardy heron server.) Hopefully someday there'll be a way to go from 32-bit to 64-bit, on the machines that can support 64-bit but had 32-bit Ubuntu installed on them.
So, last night....after making backuppc complete that day's incremental. I set out to upgrade my laptop 'ulkc' to Lucid Lynx. This is a laptop, and I originally installed Jaunty Jackalope 32-bit on it, even though its a Core 2 Duo. Main thing is that boxee didn't support 64-bit, and it was one of the key apps I wanted to run...at the time (I've never actually used it on here though...since the streams that wouldn't show on my AppleTV+Boxee also wouldn't work on here. Forcing me to resort to iTunes/AppleTV to get the content.
The upgrade was largely straight forward....though along with disabling 3rdparty sources, it removed picasa (that's the main thing I noticed missing).
Afterwards, I left the computer on, and ran a new full backup into backuppc.
This morning I did some poking around of the usual apps that I use...Firefox/Thunderbird. Of course, Thunderbird is now 3.04...so it did its upgrade/migrate thing and loss of extensions that haven't been updated for 3.x and don't work if I do 'make compatible'. I hope someday these extensions will appear....mainly the View Header toggle button and Mail Redirect.
Being that I some how am an email admin at work, needing to see full headers is something that happens regularly....and mail redirect is handy when I send a message, and then realize that I forgot to CC/BCC somebody...I can then do that with redirect and they won't see the alteration to the email that forwarding does. Also useful to 'cc' a copy of an email that was sent directly to me rather than through proper channels (even if I'll be the one that ends up doing the work).
I left the computer on again...so Thunderbird can do its new (annoying) thing....and guess it'll eventually get picked up for today's incremental...unless I go back to using it before it runs and suspend it.
Not sure when I'm going to upgrade my Ubuntu servers. 'box' is currently a karmic koala 32-bit server (older P4 that doesn't do 64-bit)...was a jaunty jackalope...though I probably should've made it a hardy heron...to match 'orac'...since there are things (namely DHCP fail-over) that I can't do due to version mismatches. But, the plan is to make both servers run the 10.04LTS...where they'll stay until 12.04LTS comes out, or not?
I'm very hesitant about touching 'orac'....because its working, has critical functions, and its temperamental. Of course, 'box' has functions more critical to my overall home network...but I'm going to risk the upgrade of it, to get it to an LTS version.
It was my original intent to not upgrade the servers until at least a couple months have past...since when I had upgraded 'box' to karmic koala shortly after its release, stuff important to me (name DNS) didn't work. apparmor was breaking it. Eventually they released the kernel upgrade that resolved the issue. Though building a 32-bit server kernel each time, etc. was also kind of painful. I've since opted to just have the server running on the 'generic-pae' kernel.
'orac' can continue to be a server, since it is a 64-bit system.
I'll probably go back to the waiting...there are other things I should get to first....
Latest Poopli Updaters -- http://lkc.me/poop
|<< <||> >>|
dsl «sans digital» orac ubuntu cfengine3 raid dvd twitter freebsd mdadm batteries «air purifier» upgrade raid1 netflix voip newegg virtualbox tivo 10.04lts b2evolution «watch instantly» ups amazon.com cox replaytv «amazon prime» tv usb «doctor who» «tivo hd» tardis backuppc zen «hd movie» ebay «windows 7» prescription eyeglasses lhaven «powersource 400» linux «windows xp» appletv boinc «chicago tardis» box «instant streaming» woot cpap