[01:59] <godfreyhk> Anyone tried to install JeOS on a non-virtual server?
[02:01] <sommer> nope... I haven't
[02:02] <godfreyhk> Do you think it'd work?
[02:04] <sommer> only if you have the same hardware as the virtual machine, jeos only includes specific drivers... to keep the weight down
[02:11] <godfreyhk> I see, thanks.
[05:59] <pleaseandthankyo> ok installed xubuntu and and goubuntu  on top of edubuntu where are they now? and how do i load them?
[06:00] <kgoetz> pleaseandthankyo: please ask in #xubuntu.
[08:14] <kraut> moin
[09:32] <juliux> ho
[09:32] <juliux> does somebody allready installed lotus notes on ubuntu? if yes is there a howto?
[09:36] <andguent> juliux: I don't know anything more than google knows. I did a basic search and found a few hits that look great.... if it was 2004
[09:36] <juliux> hehe
[10:12] <Nafallo> morning people.
[10:13] <Nafallo> if I need to deploy a vmware box. which Ubuntu should I choose? dapper has support for vmware as well until 2011?
[10:20] <_ruben> my current plans for (among other types) vmware machines is gutsy, with 'planned' upgrades to hardy when it stabilizes and then try to go lts-only from there
[10:24] <Kamping_Kaiser> Nafallo, if you install the server yes
[10:24] <Kamping_Kaiser> Nafallo, if it can wait a month, go for hardy
[10:30] <_ruben> or get out ur wallet and go for vmware esx ;)
[10:34] <\sh> Nafallo: you don't really want to use vmware-server, right?
[10:35] <\sh> esx instead is much better for commercial vmware environments...
[10:39]  * _ruben cant wait to get his esx experiments started
[10:39] <_ruben> too bad there's a few items on my todo list with higher prio :/
[11:27] <Nafallo> \sh: internal box for testing purposes. basically the intranet will be on their.
[11:27] <Nafallo> there
[11:34] <\sh> Nafallo: think about the flaws (wrong time syncs etc.) I have this problem here all the time...
[11:36] <_ruben> problem with vmware server is that dapper is probably too old, and gutsy/hardy are too new to be present on the HCL .. doesnt mean it wont work though
[11:36] <_ruben> then again, vmware's hcl updates rather slow, so dapper might even be supported ;)
[11:38] <Nafallo> \sh: I'll have a chat with you when I get back :-)
[12:27] <sommer> morning
[12:30] <nijaba> hey sommer
[12:57] <faulkes-> morning nijaba
[12:59] <nijaba> morning faulkes-
[13:01] <piedoggie> I need to put together a few virtual servers and I was wondering how well 7.10 works versus the upcoming 8.04?
[13:02] <dendrobates> piedoggie: as a guest?
[13:03] <piedoggie> guest and host
[13:03] <dendrobates> piedoggie: using what virtualisation technology?
[13:03] <piedoggie> vmware
[13:04] <dendrobates> piedoggie: both work well
[13:04] <piedoggie> sorry, I would have explained better but I am running very short on sleep
[13:05] <piedoggie> okay.  I've been playing around with the 8.04 jeos and it's not doing well.  Something fails and it doesn't get a complete sources.list file.
[13:06] <piedoggie> I guess I'll just use 7.10 and upgrade in about 30 days (ugh)
[13:06] <nijaba> piedoggie: what do you mean?  which version did you test?
[13:07] <piedoggie> I tested beta and two dailies
[13:07] <nijaba> piedoggie: I tested the jeos beta before release and I did not have a problem installing them on kvm or vmware server
[13:08] <piedoggie>  the beta would just hang on the task selection portion of the install and at roughly the same point, I would get a red screen and then the list of possible tasks.  I could load grub but I couldn't select software.  When the system booted up, I looked at sources.list and it only reference the CD-ROM
[13:08] <piedoggie> should say at roughly the same point with the dailies
[13:08] <piedoggie> if you want, I can run to a quick install now and give a more accurate report
[13:09] <nijaba> I am going to check again then...
[13:09] <piedoggie> okay, the daily image I downloaded was last night at 12:30 a.m..  Size is 101 MB
[13:10] <piedoggie> this is window so I can't give you an MD5 Jack
[13:10] <piedoggie> sorry, using speech recognition.  I need to train a new model
[13:10] <piedoggie> MD5 check
[13:12] <nijaba> piedoggie: I'll start by rechecking the beta
[13:13] <piedoggie> okay.  I'm trying to duplicate the problem with the daily from last night.  It's installing the base system now
[13:13] <piedoggie> I'm also doing this on VM workstation 6
[13:19] <piedoggie> the daily I downloaded last night has gotten further than it has with any other version I've tried
[13:20] <piedoggie> it's sitting at 90%, asking me to please wait and the virtual ethernet activity light is on solid
[13:22] <piedoggie> what is the install doing at this point?
[13:24] <nijaba> piedoggie: it could be checking the repository for update
[13:25] <piedoggie> this is probably something you can't fix now but, if a user didn't have some sort of activity indicator, they would think that the system had hung
[13:25] <piedoggie> there's a good chance you will get people thinking it's failed when really it just hasn't completed the task
[13:25] <piedoggie> on the other hand, you could always just tell them that it's a bot playing second life
[13:31] <piedoggie> finally finished doing whatever it was doing
[13:31] <piedoggie> that's a long time with no feedback
[13:32] <piedoggie> one other thing I've noticed is that the login prompt happens before you get the messages about starting the law system log daemon kernel log daemon and then local boot scripts
[13:32] <nijaba> piedoggie: 15min for an install?  does not seem that long to me.
[13:33] <nijaba> piedoggie: this is a long time issue.  You can go add your comment to the bug.  I still hope it will be fixed before final release
[13:34] <nijaba> ^ about the login prompt
[13:34] <piedoggie> I'm sorry.  I didn't pay close enough attention.  There's one phase where it's installing software I think and it sits there for something like five or more minutes without any user feedback
[13:34] <piedoggie> I'll run to the test today after I take a nap.  I really do give better error reports what I have a full nights sleep.  :-)
[13:35] <nijaba> piedoggie: np. and take care
[13:36] <piedoggie> one last question, what's the difference between vmware and kvm?  I'm becoming unhappy with vmware server and would like an alternative
[13:39] <nijaba> piedoggie: kvm is open source software, vmware is not.  kvm is a bit younger, so it does not have all the features vmware has
[13:40] <nijaba> piedoggie: but if you are running on windows, kvm will not help, it really is bound to linux, as it is part of the kernel
[13:40] <piedoggie> the only reason I'm bound to Windows is because I'm disabled and must use speech recognition
[13:41] <piedoggie> I do all of my server works with Linux although I am tempted by Solaris just because of ZFS
[13:41] <piedoggie> I have so drunk the ZFS Kool-Aid.  :-)
[13:41] <soren> I seem to missing something. What is it that is so marvelously fantastic about zfs?
[13:42] <piedoggie> ZFS is a really remarkable filesystem.  It's the first filesystem I've been excited about in almost 15 years
[13:42] <nijaba> piedoggie: hey, you do not have to excuse yourself for having window, and in your case I would be very happy if I could tell you we had some speach recog on linux, but that it not the case, AFAIK
[13:43] <piedoggie>  by erasing the boundaries between LVM, and raid, they have managed to create something that's actually better.
[13:43] <soren> piedoggie: What does it do that xfs+lvm doesn't do?
[13:43] <piedoggie> You can dynamically add or remove disks on a running system.  It is able to detect more errors to the point where they have actually detected bad power supplies through the file system
[13:43] <_ruben> xfs+lvm+raid that'd be right?
[13:44] <soren> _ruben: Possibly. lvm does some form of raid, too.
[13:44] <soren> _ruben: I've never used it, though.
[13:44] <piedoggie> it automatically migrates to other storage areas if it detects errors
[13:44] <piedoggie> the Sun zfs website has a pretty nice set of white papers on it
[13:45] <_ruben> soren: lvm only does striping afaik
[13:45] <_ruben> soren: which is a feature lacking in the debian-installer, shamefully ;)
[13:45] <soren> _ruben: You mean using more than one pv for a vg?
[13:46] <_ruben> soren: yes, and actually striping the blocks over the pv's like raid0
[13:46] <piedoggie> nijaba: with regards to speech recognition, there is some hope in that the wine people have managed to make NaturallySpeaking (the only reasonable desktop speech recognition system) work.  It runs faster and more reliably than it does on Windows (surprise surprise)
[13:46] <nijaba> piedoggie: hey cool!
[13:47] <soren> _ruben: I don't really care much about that.
[13:47] <piedoggie> but unfortunately, it doesn't speak to any Linux applications.  That is going to take some time and money and the nonprofit I belong to is working on a grant to get funding
[13:48] <soren> _ruben: a) on a reasonably busy server, the disk heads are all over the place anyway, so the disk layout of files doesn't matter much.
[13:48] <_ruben> soren: when having lots of small files it doesnt matter really .. but when storing large files, i'd say it'd be a performance improvement
[13:48] <_ruben> soren: hmm .. never looked at it that way
[13:49] <soren> _ruben: b) I like the fact that if a disk dies (and there's no redundancy), it's contained to a certain part of the filesystem rather than smashing the *Entire* thing because every other fs block is on the dead disk.
[13:49] <fromport> hmm just got bitten (again) by the hardy installer. i have > 2TB disks . setup the partions with debian-lenny boot image but during something really mingles with the partirion table and all partitions past the first 4 are wiped :-(
[13:49] <piedoggie> http://en.wikipedia.org/wiki/ZFS  is a reasonable description
[13:50] <soren> _ruben: I've never really cared about fs fragmentation on my servers. They're usually busy enough for it to not matter at all. Any place on the disk has as good a chance of being close to the disk heads at any given time as any other.
[13:50] <_ruben> soren: true, concerning b, but i'd use raid1 as pv's if its important data in any way ..
[13:50] <soren> _ruben: Sure.
[13:50] <_ruben> soren: i dont care that much about fragments either, tho increasing the number of spindles used i do prefer
[13:51] <piedoggie> thanks for the help.  I think it's time to crash and get some sleep.
[13:51] <soren> _ruben: Point a) describes why it's pointless, while point b) describes, why it's actually a bad idea.
[13:52] <soren> _ruben: Sure, but unless it's a server that's not doing anything but serving *huge* files one at a time, it won't do you much good.
[13:54] <_ruben> soren: true enough
[13:54]  * _ruben always tries to get the most out of his hardware, but that doesnt always lead to the best results/setup/whatever :p
[13:57] <_ruben> i really do need to do some research on how to get the best performance for given diskload + hardware setup
[15:03] <CrummyGummy> Hi all, is any one here using mysql clustering?
[15:10] <faulkes-> you mean ndbd and what not?
[15:10] <CrummyGummy> yup
[15:10] <faulkes-> I don't currently use it under ubuntu but I do use it
[15:11] <CrummyGummy> I'm having problems similar to this bug.
[15:11] <CrummyGummy> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=459543
[15:11] <ubotu> Debian bug 459543 in mysql-server-5.0 "mysql-server-5.0: ndb-cluster is broken" [Normal,Open]
[15:11] <CrummyGummy> And I was wondering if its valid for ubuntu.
[15:11] <juliux> does somebody get´s  domino on an ubuntu server running?
[15:17]  * CrummyGummy heads off to launchpad.
[15:24] <faulkes-> that debian bug sounds more like his cluster is borked and not actually a bug
[15:30] <CrummyGummy> Thats good and bad. I then need to find out why I have a similar problem.
[15:34] <CrummyGummy> http://pastebin.org/25478
[15:35] <CrummyGummy> faulkes-: Please have a look at that for me.
[15:35] <faulkes-> how many nodes exist in the cluster?
[15:35] <CrummyGummy> 6
[15:35] <CrummyGummy> 2 data
[15:36] <CrummyGummy> 2 mgm
[15:36] <CrummyGummy> 2 sql
[15:36] <faulkes-> have you or did you start by initializing ndbd already?
[15:37] <CrummyGummy> yes
[15:37] <faulkes-> and if this is a previously installation, do you have backups of your data as per mysql-cluster backup method?
[15:37] <CrummyGummy> Nope, this is my first time.
[15:39] <CrummyGummy> I'm pulling all the data from a multi-master replication setup.
[15:40] <faulkes-> sec
[15:47] <faulkes-> http://pastebin.org/25483
[15:48] <faulkes-> is what your error translates to
[15:48] <faulkes-> start looking for information surrounding that
[15:50] <CrummyGummy> hmmm, I've seen that error when filling the db. It shouldn't happen whne setting up a file. Quick question wrt the ndbd --initial. Does it exit by itself or do I have to stop it?
[15:50] <CrummyGummy>  /file/table/
[15:56] <CrummyGummy> cheers, gotta go...
[16:00] <faulkes-> ndbd --initial is used when you first initialize the cluster node or if you are completely rebuilding it from a backup
[16:00] <faulkes-> it acts just like a typical ndbd process afterwards
[17:12] <mathiaz> zul: re bug 204612
[17:12] <ubotu> Launchpad bug 204612 in nut "nut 2.2.1-2.1ubuntu1 fails to install on Hardy Heron" [Undecided,In progress] https://launchpad.net/bugs/204612
[17:13] <mathiaz> zul: you don't need to ask for a FFe - the package doesn't install now
[17:13] <zul> mathiaz: yep
[17:13] <mathiaz> zul: it's a bug and should be fixed
[17:13] <zul> mathiaz: ok Ill just merge it when it hits debian
[17:13] <mathiaz> zul: the problem is the udev restart call which should not be done
[17:13] <mathiaz> zul: the solution about udevtrigger is propable the RTDT
[17:14] <zul> RTDT?
[17:14] <mathiaz> zul: you may wanna ask #u-devel about this though
[17:14] <mathiaz> zul: right thing To Do
[17:14] <zul> ok will do
[17:15] <zul> mathiaz: too many acronymns
[17:15] <mathiaz> zul: yeah - especially if I start to use wrong ones...
[17:19] <zul> mathiaz: if it doesnt appear on merge-o-matic Ill just cherrypick it
[18:04] <piedoggie> how difficult is it to customize jeos with your own applications?
[18:05] <piedoggie> I should probably also say that they are all tarballs and have little or no chance of becoming .debs
[18:06] <sommer> piedoggie: should be as easy as custimizing regular ubuntu... you may just have to install more software initially, seeing as how little jeos starts out with :-)
[18:07] <piedoggie> :-)
[18:07] <piedoggie> I assume there is documentation on how to do this but I just haven't been able to find yet?
[18:08] <sommer> piedoggie: http://doc.ubuntu.com/ubuntu/serverguide/C/jeos.html is the development version of the serverguide that will be released with Hardy
[18:08] <sommer> should help get you started anyway
[18:08] <piedoggie> I may just fall back on releasing it as a virtual machine.
[18:09] <sommer> piedoggie: interesting, do you think more software vendors will take that approach in the future?
[18:10] <piedoggie> think so.  I have a customer that is a fruit and vegetable wholesaler and they are dependent on aid package called produce pro-
[18:10] <sommer> cool
[18:10] <piedoggie> the first version came on a Red Hat eight CD as a tar ball installed by the vendor
[18:10] <piedoggie> the current version comes on a virtual machine
[18:11] <piedoggie> the project I'm working on (reputation-based anti-spam) is such a bag of parts that I don't think anybody but myself can install it
[18:12] <piedoggie> so rather than inflicting a four or five page long installation process on someone, just delivered as a virtual machine and call it done
[18:12] <sommer> heh, would make sense to distribute a VM then :-)
[18:13] <jetole> hey guys
[18:13] <piedoggie> the big challenge is figuring out how to partition user data from the baseline system
[18:13] <piedoggie> I think I'm going to go with two virtual disks, one for the basic application and the other for all of the user data files
[18:14] <piedoggie> obviously I'm going to want to have as small a virtual machine as possible for distribution purposes
[18:14] <jetole> hopefully a quick question? is there a way to create a new NIC other then ifconfig eth0:1? I am not a real big fan of ifconfig but ip addr add seems to only apply an address to an existing NIC as far as I know and hping3 which I am using for testing at this point wants an interface but advanced routing seems to have some issues with eth0:1
[18:16] <piedoggie> why not add your network interface to /etc/network/interfaces
[18:23] <jetole> piedoggie: I have but it seems that ifconfig is used there however I found I can manually add an alias with ip
[18:23] <jetole> doesn't show up under ifconfig but then again neither does any additional ip on a NIC under ifconfig
[18:26] <jetole> ip link set eth6:2 up && ip addr add 192.168.0.25 dev eth6:2 ; ifconfig -a | grep '192\.168' ; ip addr list | grep '192\.168'
[18:26] <jetole> that will add the interface but will not show up under ifconfig, will show up under eth6 under ip and will allow any app to use eth6:2
[18:29] <jetole> in fact, you can lose "ip link set eth6:2 up" and just go with "ip addr add"
[19:05] <sigger> I think I broke my ubuntu server.  Will the reinstall (if there is one separate from install) preserve my etc/files, users, groups, and other things?  [apologies, I just asked this in #ubuntu, but the question is serious enough to seek 2nd opinion]
[19:16] <ScottK> sigger: No.
[19:16] <ScottK> sigger: Can you boot the server at all?
[19:17] <peterdv> (or boot from live cd and mount the partitions ?)
[19:18] <ScottK> Yep.
[19:19] <sigger> yes
[19:21] <sigger> Server boots fine.  I messed it up by cping the partition with /var then cping back.  but forgot --preserve=all so I lost all perm info
[19:21] <sigger> services not running right.  sock missing.  pid can't open. blech
[19:21] <sigger> I made a big mes
[19:22] <sigger> mess
[19:22] <sommer> sigger: do you have another "test" box you could install ubuntu server on?
[19:22] <sigger> mmmm, not really unfortunately
[19:22] <sigger> I just gave up my last "extra" box
[19:22] <sommer> you could then compare services and permissions... or you could ask per service and someone here can probably help
[19:23] <sigger> I'm not even sure why I can't ssh.  no error on server side, but no conn from client
[19:24] <sigger>  /etc/init.d/mysql start complains about not accessing .sock
[19:24] <sigger> etc
[19:25] <sommer> sigger: chown -R mysql.mysql /var/run/mysqld/ and try starting mysql
[19:25] <sigger> k
[19:29] <peterdv> Forgive me, but before attempting rescue, I feel the need to ask: "What is your backup situation". Not to be rude - just to get a feeling for the risk involved. No offence intended.
[19:31] <sigger> peterdv:  all data safely backed up off this PC.  Thank you peterdv.  No offense taken.  Helpful comment
[19:32] <sigger> k did that and perm for debian.cnf, but it now rightly complains of missing /var/run/mysqld/mysqld.sock
[19:33] <sigger> Know what that is, but no idea how to "bring it back" or get it
[19:34] <sigger> (actually there was one bit of info that I had to back up manually when the prob was discovered.  I copied the mysql data by files (with --preserve=all) rather thana mysqldump
[19:34] <sommer> sigger: the mysqld.sock file should be created when starting mysql, but if the permissions on the directory are wrong the deamon process can't create it
[19:35] <sommer> try chmod 755 /var/run/mysqld and starting mysql
[19:35] <kraut> what's the sysreq-key? (urgent)
[19:35] <kraut> alt+~?
[19:36] <sigger> actually left before the 755 msg and just tried it with 777.  no luck
[19:37] <sommer> sigger: can you pastebin the exact error?
[19:37] <sigger> starting by /etc/init.d/mysql start, rather than running mysqld directly if it matters
[19:37] <sigger> I can't because I can't even ssh in yet
[19:38] <sigger> I can retype tho.
[19:38] <sommer> sigger: does the error mention anything about perms?
[19:41] <sigger> kinda, not really, here's what I think is the relevant part:
[19:42] <sigger> Can't connect to local MySQL server through socket '/var/run/mysqld/mysqld.sock'.  Check that mysqld is running and that socket [same .sock] exists
[19:42] <sigger> a process called mysqld is running.  the .sock does not exist tho
[19:42] <sigger> hmm, lemme kill -9 and retry
[19:42] <sommer> sigger: what are the permissions and owner of /var/run/mysqld ?
[19:43] <sigger> at the moment, mysql.mysql 777
[19:44] <sommer> sigger: try chown mysql.root /var/run/mysqld, I may have told you wrong with the "-R" in the first command
[19:44] <sigger> k, I'll play with it
[19:46] <sigger> nope.  -R didn't matter anyway as theres nothing in ther
[19:47] <sommer> mmmm... what is the output of ls -l /var ?
[19:47] <sigger> err, I recall there's a bunch there
[19:47] <sigger> you want just mysql?
[19:48] <sommer> sigger: heh, yep
[19:48] <sigger> oh sorry thats not /var/run.  its /var
[19:48] <sigger> lemme get it
[19:49] <sigger> run in ls -l /var is 777 root.root
[19:49] <sigger> i.e. if i were to do ls -ld /var/run
[19:52]  * sigger starts to plan what to back up before nuking the install
[19:52] <sommer> maybe try touch /var/run/mysqld/mysqld.sock and chown it to mysql, then try starting
[19:53] <sigger> hmm good idea
[20:03] <sigger> no, but I spent some more time looking at syslog (since there is reference to it in error msg).
[20:04] <sigger> There is another error coming up where it can't lock .ibdata (i think that was the file)
[20:04] <sigger> er cannot lock ./ibdata1
[20:05] <sigger> Unable to lock ./ibdata1, error: 11
[20:05] <sommer> sigger: ah, check the perms and owner of /var/lib/mysql, should be 755 mysql.mysql
[20:06] <sigger> it is.  I confirmed above
[20:06] <sigger> well actually, lemme make it exactly that
[20:06] <sigger> its more open
[20:06] <sommer> and the contents of /var/lib/mysql should be mysql.mysql
[20:07] <sigger> contents?  you mean owner.group?
[20:07] <sigger> it was mysql.mysql 777
[20:08] <sigger> and the ibdata1 file is 777 mysql.mysql just for fun
[20:08] <sommer> sigger: yep, the contens of /var/lib/mysql should be mysql.mysql
[20:08] <sigger> well yes they all are
[20:09] <sigger> hmm, getting a little crazy here: what if I delete the lock on that file (assuming there is one)
[20:09] <sommer> sigger: worth a try, I'm running out of ideas ;)
[20:12] <sigger> nope.  I did notice that the mysqld process sticks around past a /etc/init.d/mysql stop
[20:13] <sigger> so I had to kill
[20:13] <sigger> so is it ok that I just copied my data dir files to a backup place?  i.e. can I just copy them back to newly installed mysql data dir and have everything be ok?
[20:14] <sommer> sigger: I'm not sure about that one
[20:15] <sigger> ahh maybe I can run mysqldump still
[20:16] <sigger> nope needs sock
[20:31] <zul> soren: ping https://bugs.edge.launchpad.net/ubuntu/+source/php5/+bug/204479
[20:31] <ubotu> Launchpad bug 204479 in php5 "PHP should be shipped with magic_quotes_gpc = Off  in php.ini" [Wishlist,Triaged]
[20:31] <zul> soren: what do you think?
[20:33] <faulkes-> zul: that's typical of most distributions of php.ini
[20:35] <faulkes-> I don't really think it's a huge issue, if you aren't taking steps to escape / validate your data, assuming php is going to be your magic bullet is well, going to leave you shooting blanks
[20:35] <soren> zul: I'm a bit torn on the subject.
[20:35] <owh> The PHP manual says this: http://au2.php.net/magic_quotes
[20:35] <zul> fedora has it turned off
[20:35] <owh>         magic_quotes_gpc                      Affects HTTP Request data (GET, POST, and COOKIE). Cannot be set at         runtime, and defaults to on in PHP.
[20:36] <owh> It also says that it's depreciated and removed in PHP6
[20:36] <zul> owh: I like the big fat warning there
[20:37] <owh> Personally as a PHP developer I'd expect my environment to do what the manual says. Even if it sucks.
[20:37] <owh> zul: You mean the warning :)
[20:37] <zul> yep
[20:37] <soren> There are a few troubles involved:
[20:37] <soren> magic_quotes are a more secure default.
[20:38] <soren> by extension, it teaches bad habits to web developers.
[20:39] <soren> We've had it on for a long time (don't we? (I find myself unsure)), so disabling it now, might make us vulnerable to attacks that it used to save us from.
[20:39] <faulkes-> soren: web developers? bad habits? I am shocked, *SHOCKED* to hear you even suggest something like that occurs
[20:39] <soren> It's kind of like..
[20:39]  * soren chuckles
[20:39] <soren> faulkes-: :)
[20:39] <soren> WEll, C programmers *ought* to make sure that buffer overflows didn't happen.
[20:39] <zul> soren: dunno thats why I am asking
[20:39] <owh> soren: While I agree with what you're saying, if we open that can of worms, that is, changing defaults randomly, then where does that behaviour stop?
[20:39] <zul> :)
[20:40] <soren> Yet we put the SSP stuff in place, just to be sure.
[20:40] <soren> It's not entirely comparable, though, as magic_quotes does change php's behaviour in ways the programmer will se.
[20:40] <soren> see.
[20:40] <owh> faulkes-: You should see some of the credit card code I am presented with by clients "fresh from their developer".
[20:41] <soren> owh: It sucks to change defaults, but after long consideration, I'm leaning towards "screw history. If we find better defaults, make the change."
[20:41] <owh> soren: If they were actual developers they'd know about get_magic_quotes_gpc()
[20:41] <nxvl> i can not be present on the server meeting, i feel sick and i'm going to the hospital, sorry for absense
[20:41] <owh> soren: That's a fair comment. Not sure I agree yet.
[20:41] <owh> nxvl: Good luck.
[20:41] <owh> nxvl: Get better.
[20:42] <soren> nxvl: Yeah, get better, dude!
[20:42] <owh> There is an argument for making Ubuntu better than the rest and leading the way. I totally grok that.
[20:42] <soren> owh: The sooner, the better. Usually.
[20:43] <soren> The trouble is that we're quite far along in the release cycle.
[20:43] <owh> There is another argument that says that if you are deploying a server and something as fundamental as that changes, you are pretty fsk'd.
[20:44] <owh> If, and only *if* this change is made, then there needs to be a screen explaining that on installation. It will trip up lots of people.
[20:44] <owh> Anyway, I'm a PHP developer, I'd take it either way, but then I've been doing this for a while :)
[20:45] <zul> soren: for the PHP stuff configuration file changes and memory issues we can revisit in ibex cant we?
[20:45] <owh> The level of skill out in the world is asking us for a beating.
[20:45] <owh> That is, the skill level is low and we'll be answering the support calls.
[20:46] <owh> zul: I suspect it will be PHP6 by then, but I don't know that for sure. The rules change at a major version upgrade.
[20:46] <soren> zul: I'd actually like to change the memory limits *right* now. 16 in cgi and mod_php6, and 32 in -cli.
[20:47] <zul> soren: heh if you are too busy I could do it for you of course
[20:47] <owh> soren: You mean mod_php6, or 5?
[20:47] <soren> 5.
[20:47] <soren> zul: That would be lovely!
[20:47] <owh> soren: Just to be clear, that's increasing the limit right?
[20:48] <soren> owh: I don't believe so, no.
[20:48] <owh> Are advocating making the memory limits smaller?
[20:49]  * owh checks a hardy install.
[20:49] <soren> WEll..
[20:49] <soren> I'm advocating setting the memory limits where they used to be and *ought* to be, but by accident have not been for a while.
[20:49] <owh> How long for?
[20:50] <soren> Not sure.
[20:50]  * owh checks the php manual for guidance.
[20:50] <soren> Don't.
[20:50] <soren> Well, feel free to read it, but don't be (mis)guided by it.
[20:51] <owh> "8M" before PHP 5.2.0, "16M" in PHP 5.2.0
[20:51] <owh> Currently: "128M"
[20:51] <owh> http://au2.php.net/manual/en/ini.core.php#ini.memory-limit
[20:51] <soren> owh: Close, but no cigar :)
[20:51] <owh> Uh, when I say, currently, I mean the manual. My hardy machine is still booting.
[20:52] <soren> owh: "8M" (but disabled) before PHP 5.2.0, "16M" (still disabled) in PHP 5.2.0. Now, 128MB and enabled.
[20:52] <ajmitch> oh dear, PHP
[20:52] <soren> memory_limit defaulted to *off* until 5.2.something.
[20:52] <zul> wow hey ajmitch :)
[20:52] <soren> ajmitch: Dude!
[20:52] <ajmitch> hello
[20:52] <soren> ajmitch: Long time!
[20:52] <ajmitch> not that long, honest
[20:52] <soren> :)
[20:53] <owh> soren: The shipped memory_limit is 16M
[20:53] <soren> owh: At some point the php developers figured that setting memory_limit enforcing on by default would be a good idea.
[20:53] <soren> Then they got scared that people would shout at them for breaking stuff that used to work..
[20:53] <owh> :)
[20:53] <soren> ..and then they set the default limit to 128MB to be (almost) sure that they wouldn't break anything.
[20:54] <soren> We, OTOH, have had memory_limit enforcing on for *ages*.
[20:54] <owh> So, you're cutting that back quite drastically.
[20:54] <owh> Well that's cool.
[20:54] <soren> With a default of 8MB in mod_php5 and -cgi, and 16 in -cli (to fix some pear crap, I believe)
[20:54] <owh> At the moment the php.ini file I'm looking at is 16M
[20:55] <owh> In hardy.
[20:55] <soren> Recently, the way the amount of used RAM was counted, changed. So what used to fit in 8MB didn't anymore.
[20:55] <soren> owh: Oh?
[20:55] <owh> /etc/php5/apache2/php.ini
[20:55] <nijaba> soren: I confirm it on a fresh install of php
[20:55] <soren> Did I actually manage to get around to fixing that? I'm deeply surprised. It's still in my todo list (waaaay down).
[20:55] <soren> What about -cgi and -cli?
[20:56] <owh> '/usr/share/php5/php.ini-dist.cli 32M
[20:56] <owh> '/usr/share/php5/php.ini-dist 16M
[20:57] <owh> That's all that I can see.
[20:57] <soren> O_O
[20:57] <soren> Wow.
[20:57] <owh> There are dist, paranoid and recommended ones as well.
[20:57] <soren> zul: Er... never mind, then. I (or someone else, possibly) already fixed it.
[20:57] <owh> In the examples.
[20:57] <zul> soren: yep I just checked
[20:57] <soren> dist is the template used to generate our php.ini.
[20:57] <soren> (ucf style, iirc)
[20:57] <keescook> (old topic... php magic_quotes)
[20:57] <owh> soren: Uh, the dist has 128M
[20:58] <keescook> I'd like to keep it enabled.  applications should know to deal with it -- those that don't certainly need it enabled.
[20:58] <soren> owh: You just said?
[20:58] <soren> 20:56:26 < owh> '/usr/share/php5/php.ini-dist.cli 32M
[20:58] <soren> 20:56:56 < owh> '/usr/share/php5/php.ini-dist 16M
[20:58] <soren> 20:57:02 < owh> That's all that I can see.
[20:58] <owh> Gimmie a mo.
[20:58] <owh> Yeah, the dist is in the example dir.
[20:58] <owh> '/usr/share/doc/php5-common/examples/php.ini-dist:memory_limit = 128M
[20:59] <zul>  # memory_limit: 16M for cgi/apache; 32M for cli
[20:59] <zul>         cat php.ini-dist | tr "\t" " " | sed -e'/memory_limit =/ s/\b128M/16M/g' > debian/php5-common/usr/share/php5/php.ini-dist
[20:59] <zul>         cat php.ini-dist | tr "\t" " " | sed -e'/memory_limit =/ s/\b128M/32M/g' > debian/php5-common/usr/share/php5/php.ini-dist.cli
[20:59] <owh> '/usr/share/doc/php5-common/examples/php.ini-paranoid:memory_limit = 8M
[20:59] <soren> owh: Oh. Er.. Sorry. I was being uncharacteristically ambiguous.
[20:59] <owh> '/usr/share/doc/php5-common/examples/php.ini-recommended:memory_limit = 128M
[20:59] <owh> :)
[20:59] <soren> owh: The ones you showed me first (in /usr/share/php5) are the ones used to generate our php.ini.
[21:00] <soren> owh: The reasons why are a bit involved (dpkg conffile handling stuff and such).
[21:00] <owh> Yeah, I figured, but I thought I'd be complete and then you mentioned dist, so I double checked.
[21:00]  * owh heads to meeting.
[21:43] <Saftle> have a quick question, I'm wanting to migrate all of my windows based servers into ubuntu, however I'm not sure whether ubuntu server or ubuntu desktop would be better due to me requiring a GUI. Basically we wouldl like PHP, APACHE, SSH etc to be ran from the linux environment and it would require a Virtual Machine of Windows to be running windows tasks.
[21:45] <soren> What's the gui for?
[21:45] <Saftle> to run VMware
[21:45] <soren> Ah.
[21:46] <owh> Saftle: Technically you don't actually need a gui to run vmware-server. I'm not sure if it will install without it.
[21:47] <owh> Saftle: Personally I install ubuntu-server and add what is needed.
[21:47] <Saftle> owh: so basically install gnome and vmware-server after installing ubuntu-server then?
[21:48] <Saftle> rather than using ubuntu desktop and installing lamps and vmware-server?
[21:49] <soren> The end result is *almost* identical.
[21:49] <owh> Saftle: Well, I'd not even install gnome, I'd start with vmware-server.
[21:49] <owh> Saftle: It will pull in the dependencies.
[21:50] <Saftle> owh: sorry for being a linux noob, but without the gnome GUI is vmware still visible?
[21:50] <owh> Saftle: Well, yes and no.
[21:50] <owh> Saftle: You will be able to manage the vmware-server from a remote console, but it won't be a gui on the server hardware itself.
[21:51] <Saftle> wow, ok, that might be just what I need then
[21:51] <owh> Saftle: Try it and see if it gives you what you need.
[21:52] <Saftle> owh: awesome thx alot for your help.
[21:52] <owh> Saftle: Pleasure.
[21:53] <JaxxMaxx_> To admin a VMWare'd Windows server, use RDP :)
[21:53] <owh> Saftle: Just so you know, Ubuntu is heading towards kvm. I've personally not made that switch yet, but you should be aware of it.
[21:55] <owh> JaxxMaxx_: Sometimes you need to see the BIOS screen too :)
[21:55] <Saftle> owh: you mean like a KVM switch?
[21:55] <owh> Saftle: No, like the software.
[21:56] <Saftle> owh: not familiar with it
[21:56] <JaxxMaxx_> Hmm, does VMWare have remove viewing capabilities yet?
[21:56] <Saftle> owh: just googled it, so basically it's an alternative to vmware?
[21:57] <owh> Saftle: Yes.
[21:58] <Saftle> owh: oh ok thx, I'll look into it, and see if it's a better solution than vmware
[21:58] <owh> Saftle: One difference is that the kernel support is part of Ubuntu. The packaging for vmware has sometimes been challenging.
[21:58] <JaxxMaxx_> is KVM compatible with vmware images?
[21:59] <owh> JaxxMaxx_: There are conversion tools available/in the works.
[21:59] <JaxxMaxx_> is it giving the community a tingly feeling?   why the move away from vmware?
[22:00] <JaxxMaxx_> I'd thought it was a good product, vmware
[22:00] <Saftle> owh: another noob question for you, how exactly would I go about installing xp without a gui and then be able to remote into it after installation without configuring the IP using the windows GUI? or is that all possible, but just requires a bit of googling on my end?
[22:00] <owh> JaxxMaxx_: It is, it works for me, sometimes it bites you.
[22:01] <owh> Saftle: The thing you have to bear in mind is that you're talking about a GUI into a virtual computer. A remote console.
[22:01] <owh> Saftle: You start up the hardware, insert the (virtual) CD and boot the XP installer.
[22:02] <JaxxMaxx_> I think he's just confused about how to go through the XP install without being able to see it (because the host OS has no GUI)
[22:02] <Saftle> and then from there remote into it from another computer?
[22:02] <Saftle> exactly
[22:03] <owh> Saftle: 1. Install your virtual machine layer, vmware, kvm, whatever. 2. Create a virtual machine via ssh. 3. Boot the virtual machine from a remote GUI console. 4. Install XP.
[22:04] <JaxxMaxx_> ah, like a remote X session?
[22:04] <Saftle> oh ok, that makes sense, well thx again. I'll give it a shot. I'm sure that doing it this way will definately decrease the worlkload on the server
[22:04] <Saftle> workload*
[22:10] <owh> Saftle: Well, the gui doesn't draw that much cycles if it isn't doing anything.
[22:11] <JaxxMaxx_> Just occupies RAM
[22:25] <owh> I'm trying to download stuff from bazaar.launchpad.net and I'm getting 500 Internal Errors. Any ideas?
[22:25] <zylstra555> Hello. I am using Ubuntu server, I get this when running apt-get install: E: Could not get lock /var/cache/apt/archives/lock - open (11 Resource temporarily unavailable) | E: Unable to lock the download directory
[22:27] <owh> zylstra555: Are you running the apt-get as root, is another package installer already running?
[22:28] <zylstra555> owh: Yes, Apt-get is in root. I am getting help in #ubuntu at the moment... will return here if they cant come up with anything. Also, its only one instance of apt-get running
[22:29] <owh> zylstra555: What about aptitude, synaptic, dpkg?
[22:30] <zylstra555> sudo fuser -vki /var/lib/dpkg/lock;sudo dpkg --configure -a   fixed the problem, so, no idea what caused it
[22:30] <owh> zylstra555: Is this in hardy?
[22:30] <zylstra555> owh: Shoot, forgot I had it installed. Yes, it is, in fact
[22:31] <zylstra555> (I should write the OS version on the side of it... should have mentioned that immediately)
[22:36] <zylstra555> Would it be a security risks if I made my regular user account into a full administrator without need of using the SUDO command?
[22:37] <owh> zylstra555: In short, yes.
[22:37] <owh> zylstra555: If you can reproduce this problem, can you please report a bug?
[22:38] <zylstra555> owh: Yes, however, I dont think I can. I have no idea what caused it. I did do something evil and decide to install Samba over Webmin, which, uses the -F option
[22:38] <owh> ROTFL, that will learn you :)
[22:39] <owh> zylstra555: I wonder if the samba package was still installing.
[22:39] <zylstra555> owh: It pretty much stopped. I restarted and tried again, and restarted and tried a different package, the same problem happened.
[22:40] <owh> zylstra555: What I'm saying is that perhaps webmin was still trying to install.
[22:40] <zylstra555> owh: Hmm... perhaps
[22:43] <zylstra555> Okay, second question: I am having trouble getting data into my public_html folder, located at /home/user/public_html  and I cant use my FTP client to change the attributes, since it doesnt have the proper privileges (thus, why I asked about making my regular account into a full root)
[22:44] <zylstra555> How do I change it? I tried sudo chmod +r+w public_html   but that didnt yield anything, telling me I properly used the wrong switches.
[22:45] <mok0> zylstra555: chmod 751 ~/public_html
[22:46] <mok0> zylstra555: and make sure all .html files have protection 644
[22:46] <zylstra555> 664? Write, Read, but not execute?
[22:47] <zylstra555> No, I take that back
[22:47] <mok0> zylstra555: yes, owner rw others r
[22:47] <zylstra555> I need my webserver to be executable, and certain parts to be writtable. (Which, I can deal with later)
[22:47] <mok0> zylstra555: 751 = owner rwx, group rx, others x
[22:48] <mok0> zylstra555: writable? eeech.
[22:48] <zylstra555> mok0: Forum :/
[22:48] <mok0> zylstra555: make the directory or files belong to group www-data
[22:48] <zylstra555> mok0: It does
[22:48] <mok0> and give group write access
[22:49] <zylstra555> (Default is /var/www/ for Apache2, I changed it to /home/user/public_html for some extremely strange reason. )
[22:49] <mok0> i.e 664 for files and 771 for dirs
[22:49] <mok0> zylstra555: heh, well that's up to you :-)
[22:49] <zylstra555> I just cant get chmod 751 ~/public_html to work at all
[22:50] <mok0> zylstra555: do you own it?
[22:50] <zylstra555> It keeps saying that public_html Does not exist
[22:50] <zylstra555> mok0: My server is screwy.. other than that, it belongs to the www-data group, its not necessarily part of mine. How would I change that?
[22:50] <mok0> what does "id" say?
[22:51] <mok0> zylstra555: you don't
[22:51] <zylstra555> (well, not associated with my user account, that is. I am the administrator... )
[22:52] <zylstra555> ID:    gid=1000(server) groups=4(adm),20(dialout),24(cdrom),2(floppy),29(audio),30(dip),44(video),46(plugdev),104(scanner),10(lpadmin),110(admin),1000(server)
[22:52] <mok0> zylstra555: go "id" in the terminal and paste the result
[22:52] <mok0> heh
[22:52] <mok0> ok
[22:52] <zylstra555> Oh, I dont know if this has anything to do with it, but, the command "su" does not work for me
[22:52] <mok0> zylstra555: grep server /etc/passwd
[22:52] <mok0> zylstra555: right
[22:52] <Bidou> 'night all
[22:52] <zylstra555> server:x:1000:1000:server,,,:/home/server:/bin/bash
[22:53] <zylstra555> su returns: su: Authentication failure
[22:53] <mok0> zylstra555: ok, so ~/public_html refers to /home/server/public_html
[22:53] <mok0> zylstra555: ~ means -> users home dir
[22:53] <zylstra555> mok0: So, now it should just give me the same permissions as my current home folder, right?
[22:54] <mok0> zylstra555: don't know what you mean, exactly
[22:55] <zylstra555> mok0: I still cant set file attributes for the public_html folder
[22:55] <mok0> zylstra555: does it exist?
[22:55] <zylstra555> mok0: Yes, quite plain and clearly on the DIR command
[22:55] <mok0> ls -l ~/public_html
[22:55] <zylstra555> in fact, cd public_html works fine
[22:55] <zylstra555> It keeps saying:
[22:55] <zylstra555> "No such file or directory"
[22:55] <mok0> ls -l ~/public_html
[22:56] <zylstra555> The same thing happens
[22:56] <zylstra555> Should I stop Apache2 and delete the folder and create it again?
[22:56] <mok0> no such file or directory?
[22:56] <mok0> zylstra555: NP
[22:56] <mok0> zylstra555:  NO
[22:56] <zylstra555> ls: /home/server/public_html: No such file or directory
[22:56] <zylstra555> wait, /server/?
[22:56] <mok0> zylstra555: then: mkdir ~/public_html
[22:58] <zylstra555> ls -l ~/public_html returns: "total 0"
[22:58] <mok0> ok, try ls -ld ~/public_html
[22:58] <zylstra555> Its kind of funny, I had my server running fine before... then I decided to redo everything
[22:58] <zylstra555> drwxr-xr-x 2 server server 4096 2008-03-26 15:56 /home/server/public_html
[22:58] <mok0> zylstra555: the server has nothing to do with this.
[22:59] <mok0> zylstra555: ok, now your ~/public_html dir looks fine
[22:59] <mok0> zylstra555: when you said "it worked fine" was that before changing away from /var/www ?
[23:00] <zylstra555> mok0: I still cant upload to it via FTP, thus, the file attributes are still fine
[23:00] <mok0> zylstra555: hmm
[23:01] <mok0> is this forum a home-made one, or what?
[23:01] <zylstra555> mok0: Nope, my server was only allocating 3GB of 20 that I had available. This was due to using the DD command for moving the drives data. I decided to do an entire reinstall so I could fix most of the other problems it was having anyways. Before I had the public_Html folder in /home/user/public_html and I decided to set it back to the same one
[23:01] <zylstra555> mok0: The forum (which is not currently on the server, since I cannot upload it) is a SMF forum. The server itself is sitting behind me
[23:02] <mok0> zylstra555: I guess first you should get the httpd server to display a simple html page
[23:03] <zylstra555> mok0: I cant. I cant do anything to the public_html directory, let alone create an index.html page
[23:03] <mok0> zylstra555: but you need to tell apache2 that there is another doc dir
[23:03] <mok0> zylstra555: still not?
[23:03] <mok0> zylstra555: that's very puzzling
[23:03] <zylstra555> mok0: I still cant do anything to that folder, no, if thats what you were asking. (The irony... I dont know how I got it working before)
[23:04] <mok0> zylstra555: above, you have a directory ~/public_html that looks quite normal
[23:05] <mok0> zylstra555: in the terminal, try "cd ~/public_html"
[23:05] <zylstra555> (currently in the /home/user/ folder)
[23:05] <zylstra555> mok0: cd ~/public_html works fine
[23:05] <mok0> zylstra555: now, "touch index.html"
[23:06] <mok0> zylstra555: that should give you an empty file
[23:06] <zylstra555> mok0: Okay
[23:06] <mok0> zylstra555: what editor are you comfortable with?
[23:06] <zylstra555> mok0: that worked
[23:06] <zylstra555> nano,
[23:06] <zylstra555> mok0: Creating a very basic page
[23:06] <mok0> zylstra555: ok, then try writing something in that empty file using nano
[23:07] <zylstra555> mok0: I can write to that file
[23:07] <mok0> yay
[23:08] <zylstra555> When, on my local network, I visit http://192.168.1.7/ I get the Apache Directory listing, which, is blank
[23:08] <zylstra555>  Also, my FTP client doesnt list the index.html file
[23:09] <mok0> zylstra555: very slow access from here
[23:09] <mok0> zylstra555: try http://192.168.1.7/~server
[23:09] <zylstra555> mok0: Are you trying to visit it?
[23:09] <mok0> zylstra555: yes
[23:10] <zylstra555> mok0: 192.168.1.X isnt a regular IP address, thats a router assigned address. Right now, my site is at www.zylstrablog.co.nr and is being hosted by my Windows computer
[23:10] <zylstra555> (which, also, happens to be down at the moment..
[23:10] <mok0> zylstra555: I see
[23:11] <zylstra555> mok0: I see where we were on two pages. The directory /home/server/public_html now exists... but, what about /home/user/public_htm ?
[23:11] <mok0> zylstra555: anyway, since you put the index.html page in the user "server" 's public_html directory, you need to address it with ~server after the url
[23:11] <zylstra555> *html
[23:11] <zylstra555> http://192.168.1.7/~server returns "NOT FOUND"
[23:11] <mok0> zylstra555: /home/user/public_html would belong to a user called "user"
[23:12] <mok0> zylstra555: you probably have to tune apache's conf files
[23:12] <mok0> zylstra555: hang on
[23:12] <zylstra555> mok0: Correct. Whome, does not exist. (This is how it used to be configured)    Apache is set to use /home/user/public_html
[23:13] <zylstra555> www-data's home folder is /home/user/public_html
[23:13] <mok0> zylstra555: normally, you don't do it like that.
[23:13] <zylstra555> mok0: :)
[23:13] <mok0> zylstra555: apache can serve any directory
[23:14] <mok0> zylstra555: ordinary users can serve their ~/public_html directory
[23:14] <mok0> zylstra555: will you have users on the machine?
[23:14] <zylstra555> mok0: Nope, no other users
[23:15] <mok0> zylstra555: then you might as well edit apache's conf files to tell it where the document root is
[23:15] <mok0> zylstra555: it doesn't have to be a home directory for anyone
[23:15] <zylstra555> mok0: Which, is currently set to /home/user/public_html    should I set it to /home/server/public_html ?
[23:15] <mok0> zylstra555: by default, it is /var/www
[23:17] <mok0> zylstra555: look in /etc/apache2/sites-enabled
[23:18] <zylstra555> mok0: I have to leave my computer for a while, if you happen to be here when I get back (which, might take about 30 minutes or so) I will be able to continue. I greatly thank you for your help, hopefully, I will get this resolved
[23:18] <mok0> It is quite late here, so I might be gone
[23:20] <mok0> Anyway, I advice you to edit the  apache config file, look for "DocumentRoot": that will be where the basic URL goes to
[23:24] <Erickj92> does anyone have some time to help me trouble shoot apache2?