The Linux Questions Thread: a bunch of pitfalls, but technically it's possible

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > The Linux Questions Thread: a bunch of pitfalls, but technically it's possible

«‹›1029 »

Saukkis: May 16, 2003; Unless I'm on the inside curve pointing straight at oncoming traffic the high beams stay on and I laugh at your puny protest flashes.
I am Most Important Man. Most Important Man in the World.

Has anyone tried using Ubuntu with a wide vertical Gnome panel on the side? It seems to be pretty broken. Below is a picture of how it works and under it an example of how I would want it to work from Vista. Not only does it get completely messed up even switching between applications doesn't work right. It often required repeated mouseclicks before it registers.

Click here for the full 1440x900 image.

Click here for the full 959x600 image.

# ? Jan 24, 2010 19:13

Adbot: ADBOT LOVES YOU

# ? May 10, 2024 17:47

flyboi: Oct 13, 2005; agg stop posting; College Slice

edit: im an idiot.

for those that saw the issue here's the final results. nagios plugin to check for read-only on a given test location

code:

#!/bin/bash

STATE_OK=0
STATE_CRITICAL=2
OUTPUT=`touch ${1} 2>&1`

if [[ `printf "%s" "$OUTPUT" | grep "Read-only"` ]]
then
        echo "ERROR ${1} IS READ ONLY"
        exit $STATE_CRITICAL
else
        echo "${1} IS WRITABLE"
        exit $STATE_OK
fi

flyboi fucked around with this message at 16:26 on Jan 27, 2010

# ? Jan 27, 2010 16:17

m5: Oct 1, 2001

flyboi posted:

I'm trying to write a script that will check if a file system is read only by touching it.

A basic version of what I've got:
code:
#!/bin/bash

OUTPUT=`touch /tmp/test 2>&1`

if [ `printf "%s" "$OUTPUT" | grep "Read-only"` ]
then
   echo "ERROR FILESYSTEM IS READ-ONLY"
else
   echo "FILE SYSTEM IS WRITABLE"
fi
But every time the script is run the stderr from the touch always outputs to the terminal rather than to the variable. This causes the if case to say the file system is writable. Ideas?

Use "grep -q".

# ? Jan 27, 2010 16:22

m5: Oct 1, 2001

Also it's cheaper to use the shell's built-in pattern matching to check the error. The "case" statement gives you access to that, albeit in a weird way.

# ? Jan 27, 2010 16:24

flyboi: Oct 13, 2005; agg stop posting; College Slice

Yeah a case would be less work but I want it to be readable so the people who work under me can understand what it does. A few of them are trying to learn bash and the sort and like checking out my scripts. Here's a cool one I wrote to check the health on 3ware raid cards for nagios:

code:

#!/bin/bash
 
# Exit codes for nagios
STATE_OK=0
STATE_CRITICAL=2
 
# Grab available controllers
controller=`tw_cli info|grep -E "^c"|awk '{print $1}'`
 
for e in $controller
do
        # Grab number of raid arrays and ports on available controller
        raids=`tw_cli info $e | awk '{print $1}' | grep u[0-9]`
        ports=`tw_cli info $e | grep -v "NOT-PRESENT" | awk '{print $1}' | grep p[0-9]`
 
        # Checks raid status
        for i in $raids
        do
                status=`tw_cli info $e $i status | awk '{print $4}'`
                case "$status" in
                        OK)
                                ;;
                        *)
                                message="$message ${i} on $e is $status;"
                                ;;
                esac
        done
 
        #Checks port status but considers not-present unimportant
        for i in $ports
        do
                status=`tw_cli info $e $i status | awk '{print $4}'`
                case "$status" in
                        OK)
                                ;;
                        *)
                                message="$message ${i} on $e is $status;"
                                ;;
                esac
        done
done
 
# Do we have an error?
if [[ ${message} == "" ]]
then
        echo Raid array is OK
        exit $STATE_OK;
else
        echo CRITICAL - ${message}
        exit $STATE_CRITICAL;
fi

# ? Jan 27, 2010 16:38

Carthag Tuek: Oct 15, 2005; Tider skal komme,
tider skal henrulle,
sl�gt skal f�lge sl�gters gang

Speaking of Nagios, I have a bunch of nfs mounts that are each mounted on different subsets of machines. Right now I'm checking free space by checking drives a, b, & c on app1 and drives c, d, & e on dev3 and so on and as you can tell that's a mess. All of this via NRPE:

command[check_drive_a]=/usr/lib/nagios/plugins/check_disk -w 40% -c 60% -p "/mnt/drive_a"

But I guess no clean way other than to have some machine somewhere with everything mounted?

# ? Jan 27, 2010 16:47

flyboi: Oct 13, 2005; agg stop posting; College Slice

You can enable arguments via the nrpe config and change the check to be something like:
(on nagios server)
check_nrpe -h $HOSTADDRESS -c check_disk -a '-w 5% -c2% -p /mnt/drive_a'

(on nrpe)
command[check_drive]=/usr/lib/nagios/plugins/check_disk $ARG1$

That way you can have nrpe more uniform and leave the command delegation to nagios. You also need to make sure nrpe is only configured to listen to your nagios server as arguments can be a security risk.

# ? Jan 27, 2010 17:03

Carthag Tuek: Oct 15, 2005; Tider skal komme,
tider skal henrulle,
sl�gt skal f�lge sl�gters gang

yeah I was looking a little into that, how bad the security would be.

All the hosts except the front-ends are behind a firewall, nrpe is already configured to only accept connections from the nagios host, so enabling arguments probably wouldn't be that risky.

# ? Jan 27, 2010 19:55

Bob Morales: Aug 18, 2006; ~~Just wear the fucking mask, Bob~~

I don't care how many people I probably infected with COVID-19 while refusing to wear a mask, my comfort is far more important than the health and safety of everyone around me!

I wish there were good IRC channels for Linux around.

Freenode is basically ran by dicks and they have so many goddamn rules. I understand it can't be a free for all but they're just draconian.

EFnet used to be great, but it's dead. You'll have a hundred people in a channel and literally every single one is an idler.

:smith:

# ? Jan 27, 2010 22:29

covener: Jan 10, 2004; You know, for kids!

Bob Morales posted:

Freenode is basically ran by dicks and they have so many goddamn rules. I understand it can't be a free for all but they're just draconian.

Like what? When people volunteer their valuable time to help others, they don't want that time wasted by people who haven't done their homework and can't perform simple problem determination.

# ? Jan 27, 2010 22:52

bitprophet: Jul 22, 2004; Taco Defender

Bob Morales posted:

Freenode is basically ran by dicks and they have so many goddamn rules. I understand it can't be a free for all but they're just draconian.

How so? I mean, yea, the network itself seems very Debian-esque in its policy and organization, but I don't see how that limits the channels' ability to hold conversations and such. Edit: gently caress, beaten, and also: did you mean the Linux channels on Freenode, or Freenode itself?

# ? Jan 27, 2010 22:54

Vulture Culture: Jul 14, 2003; I was never enjoying it. I only eat it for the nutrients.

About 100% of the chanops on FreeNode have a problem with my username, for one. They're humorless sacks of poo poo.

# ? Jan 27, 2010 23:03

Peanutmonger: Dec 6, 2002

flyboi posted:

<script>

Bash has much easier ways to test writability, among other things. See "help test" to see various tests that can be performed by [ (also known as test, see "man test"). For example:

code:

#!/bin/bash

STATE_OK=0
STATE_CRITICAL=2

if [ -w "${1}" ]
then
        echo "${1} IS WRITABLE"
        exit $STATE_OK
else
        echo "ERROR ${1} IS READ ONLY"
        exit $STATE_CRITICAL
fi

Peanutmonger fucked around with this message at 03:19 on Jan 28, 2010

# ? Jan 28, 2010 03:07

JHVH-1: Jun 28, 2002

Misogynist posted:

About 100% of the chanops on FreeNode have a problem with my username, for one. They're humorless sacks of poo poo.

Takes one to know one I guess.

I have had my nick registered there for over 10 years. Never had a problem.

# ? Jan 28, 2010 11:53

Bob Morales: Aug 18, 2006; ~~Just wear the fucking mask, Bob~~

I don't care how many people I probably infected with COVID-19 while refusing to wear a mask, my comfort is far more important than the health and safety of everyone around me!

bitprophet posted:

How so? I mean, yea, the network itself seems very Debian-esque in its policy and organization, but I don't see how that limits the channels' ability to hold conversations and such. Edit: gently caress, beaten, and also: did you mean the Linux channels on Freenode, or Freenode itself?

Linux channels

It's almost like they avoid helping people on purpose. TRY CHANNEL #someotherchannel. WE DONT SUPPORT THAT. THAT PACKAGE ISN'T RECOMMENDED TO BE USED. COME BACK TO US AFTER YOU READ THIS LINK <link>. Especially if it's a very simple question.

I understand that you need to research the problem on your own, and hate people that sit in a channel and ask the simplest questions over and over again just as much as anyone else. But these guys will literally interrupt your conversation with someone and threaten to kick you out of the channel if you don't shut up.

# ? Jan 28, 2010 15:03

Vulture Culture: Jul 14, 2003; I was never enjoying it. I only eat it for the nutrients.

Hey, I've got a weird shared object issue with a locally-built SQLite 3.6 on one of my CentOS 5 servers. (CentOS ships SQLite 3.3.)

I'll summarize as follows:

There shouldn't be any problems linking the SO to any of its dependencies:

code:

root@server:~# ldd /opt/sqlite-3.6.16/lib/libsqlite3.so.0
        libncurses.so.5 => /usr/lib64/libncurses.so.5 (0x00002b8385843000)
        libpthread.so.0 => /lib64/libpthread.so.0 (0x00002b8385aa0000)
        libc.so.6 => /lib64/libc.so.6 (0x00002b8385cbb000)
        libdl.so.2 => /lib64/libdl.so.2 (0x00002b8386013000)
        /lib64/ld-linux-x86-64.so.2 (0x00000035c4400000)

And yet:

code:

root@server:~# cat /etc/ld.so.conf.d/local-sqlite.conf
/opt/sqlite-3.6.16/lib <---
root@server:~# ldconfig -p | grep sqlite
        libsqlite3.so.0 (libc6,x86-64) => /opt/sqlite-3.6.16/lib/libsqlite3.so.0 <---
        libsqlite3.so.0 (libc6,x86-64) => /usr/lib64/libsqlite3.so.0
        libsqlite3.so.0 (libc6) => /usr/lib/libsqlite3.so.0
        libsqlite3.so (libc6,x86-64) => /opt/sqlite-3.6.16/lib/libsqlite3.so
root@server:~# ldd /usr/lib64/python2.6/lib-dynload/_sqlite3.so
        libsqlite3.so.0 => /usr/lib64/libsqlite3.so.0 (0x00002b94978ef000) <---
        libpython2.6.so.1.0 => /usr/lib64/libpython2.6.so.1.0 (0x00002b9497b4a000)
        libpthread.so.0 => /lib64/libpthread.so.0 (0x00002b9497efe000)
        libc.so.6 => /lib64/libc.so.6 (0x00002b949811a000)
        libdl.so.2 => /lib64/libdl.so.2 (0x00002b9498471000)
        libutil.so.1 => /lib64/libutil.so.1 (0x00002b9498675000)
        libm.so.6 => /lib64/libm.so.6 (0x00002b9498879000)
        /lib64/ld-linux-x86-64.so.2 (0x00000035c4400000)
root@server:~# LD_PRELOAD=/opt/sqlite-3.6.16/lib/libsqlite3.so.0 ldd /usr/lib64/python2.6/lib-dynload/_sqlite3.so
        /opt/sqlite-3.6.16/lib/libsqlite3.so.0 (0x00002ab6038fe000) <---
        libpython2.6.so.1.0 => /usr/lib64/libpython2.6.so.1.0 (0x00002ab603b6a000)
        libpthread.so.0 => /lib64/libpthread.so.0 (0x00002ab603f1e000)
        libc.so.6 => /lib64/libc.so.6 (0x00002ab60413a000)
        libncurses.so.5 => /usr/lib64/libncurses.so.5 (0x00002ab604491000)
        libdl.so.2 => /lib64/libdl.so.2 (0x00002ab6046ee000)
        libutil.so.1 => /lib64/libutil.so.1 (0x00002ab6048f3000)
        libm.so.6 => /lib64/libm.so.6 (0x00002ab604af6000)
        /lib64/ld-linux-x86-64.so.2 (0x00000035c4400000)

---

root@server:~# python26
Python 2.6.4 (r264:75706, Jan 11 2010, 11:16:55)
[GCC 4.1.2 20080704 (Red Hat 4.1.2-46)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import sqlite3
>>> sqlite3.sqlite_version
'3.3.6' <---
>>>
root@server:~# LD_LIBRARY_PATH=/opt/sqlite-3.6.16/lib python26
Python 2.6.4 (r264:75706, Jan 11 2010, 11:16:55)
[GCC 4.1.2 20080704 (Red Hat 4.1.2-46)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import sqlite3
>>> sqlite3.sqlite_version
'3.3.6' <---
>>>
root@server:~# LD_PRELOAD=/opt/sqlite-3.6.16/lib/libsqlite3.so.0 python26
Python 2.6.4 (r264:75706, Jan 11 2010, 11:16:55)
[GCC 4.1.2 20080704 (Red Hat 4.1.2-46)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import sqlite3
>>> sqlite3.sqlite_version
'3.6.16' <---
>>>

Any idea why this isn't loading this shared library unless it's explicitly specified in LD_PRELOAD?

Vulture Culture fucked around with this message at 18:03 on Jan 28, 2010

# ? Jan 28, 2010 16:36

Bob Morales: Aug 18, 2006; ~~Just wear the fucking mask, Bob~~

I don't care how many people I probably infected with COVID-19 while refusing to wear a mask, my comfort is far more important than the health and safety of everyone around me!

Here's where someone from Freenode would tell you to use the SQLite packages from CentOS, or send you to #sqlite

# ? Jan 28, 2010 17:49

covener: Jan 10, 2004; You know, for kids!

Misogynist posted:

Any idea why this isn't loading this shared library unless it's explicitly specified in LD_PRELOAD?

ENOCLUE, but LD_DEBUG=all and LD_DEBUG_FILE=/tmp/ld.out will keep you busy for a while in the LD_LIBRARY_PATH case.

# ? Jan 28, 2010 19:20

Vulture Culture: Jul 14, 2003; I was never enjoying it. I only eat it for the nutrients.

covener posted:

ENOCLUE, but LD_DEBUG=all and LD_DEBUG_FILE=/tmp/ld.out will keep you busy for a while in the LD_LIBRARY_PATH case.

Thanks. I found this:

code:

     29702:     file=/usr/lib64/python2.6/lib-dynload/_sqlite3.so [0];  needed by /usr/lib64/libpython2.6.so.1.0 [0]                    
     29702:     file=/usr/lib64/python2.6/lib-dynload/_sqlite3.so [0];  generating link map                                             
     29702:       dynamic: 0x00002b301d92eac8  base: 0x00002b301d71f000   size: 0x0000000000211df0                                      
     29702:         entry: 0x00002b301d725cc0  phdr: 0x00002b301d71f040  phnum:                  5                                      
     29702:                                                                                                                             
     29702:                                                                                                                             
     29702:     file=libsqlite3.so.0 [0];  needed by /usr/lib64/python2.6/lib-dynload/_sqlite3.so [0]                                   
     29702:     find library=libsqlite3.so.0 [0]; searching                                                                             
     29702:      search path=/usr/lib64/tls/x86_64:/usr/lib64/tls:/usr/lib64/x86_64:/usr/lib64          (system search path)            
     29702:       trying file=/usr/lib64/tls/x86_64/libsqlite3.so.0                                                                     
     29702:       trying file=/usr/lib64/tls/libsqlite3.so.0                                                                            
     29702:       trying file=/usr/lib64/x86_64/libsqlite3.so.0                                                                         
     29702:       trying file=/usr/lib64/libsqlite3.so.0

Now I'm trying to decipher why it's not looking at all where ldconfig says it should look.

# ? Jan 28, 2010 20:01

spiritual bypass: Feb 19, 2008; Grimey Drawer

I'd like to run a program every 10 seconds, but crontab only seems to have space for jobs down to the minute.

Is there a relatively simple way to get that working or should I just suck it up and lower my frequency to one minute?

# ? Jan 28, 2010 23:52

ShoulderDaemon: Oct 9, 2003; support goon fund; Taco Defender

rt4 posted:

I'd like to run a program every 10 seconds, but crontab only seems to have space for jobs down to the minute.

Is there a relatively simple way to get that working or should I just suck it up and lower my frequency to one minute?

Normally for tasks like that you'd create a separate daemon to manage the task, perhaps as simple as a shell script that runs it in a loop with a delay. Typically cron isn't appropriate for very-frequent tasks because you have subtle other requirements like "don't run more than one instance at a time" and "stop trying to run it if it fails because I don't want 3000 emails with the same error message in my inbox in the morning", and cron is not good at managing those constraints.

But, more practically, 10 seconds is a really short delay. What does this task do, and why does it need to run so often? It's likely there's a better way to do what you want that doesn't involve polling so frequently.

# ? Jan 28, 2010 23:57

spiritual bypass: Feb 19, 2008; Grimey Drawer

ShoulderDaemon posted:

But, more practically, 10 seconds is a really short delay. What does this task do, and why does it need to run so often? It's likely there's a better way to do what you want that doesn't involve polling so frequently.

Business reasons :rolleye:

It's easier to program stupid bullshit than it is to convince them that it's not going to have the desired effect. Polling at 10 seconds isn't going to cause any problems in this instance. I understand that there's probably a cleaner solution, but this one requires the fewest lines of code and my morale could not get much lower.

I think I'll just make a shell script that runs the program, waits 10 seconds, then iterates 5 more times. Thanks for the advice.

# ? Jan 29, 2010 00:09

Rastor: Jun 2, 2001

Edit: nevermind

# ? Jan 29, 2010 03:54

Bushwack: Aug 29, 2000

The "watch" command will do what you want.

# ? Jan 29, 2010 04:39

DizzyBum: Apr 16, 2007

What's the usual method for permanently blocking an IP address from a server? Adding it to iptables or the kernel routing table is great but I need it to persist after a reboot.

# ? Jan 29, 2010 13:02

JHVH-1: Jun 28, 2002

DizzyBum posted:

What's the usual method for permanently blocking an IP address from a server? Adding it to iptables or the kernel routing table is great but I need it to persist after a reboot.

A distro like redhat/centos has default iptables rules /etc/sysconfig/iptables that you can add it to. If your distro doesn't use that, then you can stick a file with rules in /etc/iptables.rules or whatever you want to call it. Just run iptables-save /etc/iptables.rules. Then put 'iptables-restore /etc/iptables.rules' in /etc/rc.local or make a proper rc script.

Basically read your distro's documentation to find out how they expect to do it and set it up.

# ? Jan 29, 2010 13:51

DizzyBum: Apr 16, 2007

JHVH-1 posted:

A distro like redhat/centos has default iptables rules /etc/sysconfig/iptables that you can add it to. If your distro doesn't use that, then you can stick a file with rules in /etc/iptables.rules or whatever you want to call it. Just run iptables-save /etc/iptables.rules. Then put 'iptables-restore /etc/iptables.rules' in /etc/rc.local or make a proper rc script.

Basically read your distro's documentation to find out how they expect to do it and set it up.

That's handy. Thanks!

# ? Jan 29, 2010 14:02

other people: Jun 27, 2004; Associate Christ

Here is my fstab line for my compact flash drive:

/dev/sdd1 /mnt/cflash auto noauto,rw,user,uid=patrick,gid=users,dmask=0022,fmask=0133 0 0

When I mounted this, all the file names used to be lower case. Now, for some reason all file names are upper case. I guess debian updated something that changed this?

I have added shortname=lower to the options, which works, but I am not 100% sure this won't cause me trouble. Can any one confirm/deny that it might be an issue?

And what did debian change? google isn't helping here.

# ? Feb 2, 2010 05:01

Weatherman: Jul 30, 2003; WARBLEKLONK

Is anyone using or aware of some software package that I can use on my server as an alternative to RapidShare and the like?

My office occasionally needs to transfer large DTP files to another branch within the country, and a contractor outside the country. Since we're just a small group with infrequent needs, the person with the file usually just uses one of the free services. Being ex-IT staff myself, I'm a little wary of throwing some of our work in progress on servers residing in who knows where run by who knows.

I have a FTTH connection with a server running Ubuntu 9.10 with 240 gig of storage attached to it. Neither my home connection nor the office are subject to data upload/download restrictions so that's not a problem, and my server isn't doing much anyway. I'd like to run something that presents a simple upload interface on a web page, takes the file, bungs it on the hard drive and generates a link that the file can be downloaded from.

Unfortunately I'm not familiar with the search terms I could use to narrow down the results Google is spewing at me, so I'd like to know if you have a recommended package.

# ? Feb 2, 2010 07:29

GringoGrande: Jul 27, 2001; Nah...

Weatherman posted:

Is anyone using or aware of some software package that I can use on my server as an alternative to RapidShare and the like?

I haven't tried it, but maybe quickserve might work?

# ? Feb 2, 2010 09:00

dont skimp on the shrimp: Apr 23, 2008

Kaluza-Klein posted:

Here is my fstab line for my compact flash drive:

/dev/sdd1 /mnt/cflash auto noauto,rw,user,uid=patrick,gid=users,dmask=0022,fmask=0133 0 0

When I mounted this, all the file names used to be lower case. Now, for some reason all file names are upper case. I guess debian updated something that changed this?

I have added shortname=lower to the options, which works, but I am not 100% sure this won't cause me trouble. Can any one confirm/deny that it might be an issue?

And what did debian change? google isn't helping here.

Sounds like you're using FAT32 on the stick, my guess is they updated dosfstools.

FAT32 isn't case sensitive, so it shouldn't cause any trouble at all.

# ? Feb 2, 2010 11:25

Bob Morales: Aug 18, 2006; ~~Just wear the fucking mask, Bob~~

I don't care how many people I probably infected with COVID-19 while refusing to wear a mask, my comfort is far more important than the health and safety of everyone around me!

Weatherman posted:

Is anyone using or aware of some software package that I can use on my server as an alternative to RapidShare and the like?

My office occasionally needs to transfer large DTP files to another branch within the country, and a contractor outside the country. Since we're just a small group with infrequent needs, the person with the file usually just uses one of the free services. Being ex-IT staff myself, I'm a little wary of throwing some of our work in progress on servers residing in who knows where run by who knows.

# ? Feb 2, 2010 14:17

Carthag Tuek: Oct 15, 2005; Tider skal komme,
tider skal henrulle,
sl�gt skal f�lge sl�gters gang

On Debian, what's the correct way to do this?

The hosts are configured with static IPs in /etc/network/interfaces

Currently, /etc/hosts has the following:

127.0.0.1 localhost
127.0.1.1 hostname.domain.tld hostname

However, I've read that I should use the following:

127.0.0.1 localhost
w.x.y.z hostname.domain.tld hostname

Where w.x.y.z is the static IP.

Seems if I have one, some processes fail, if I have the other, other processes fail. (for instance some of the hosts run tomcat and require the real IP, but wotaskd (WebObjects) stopped working unless I had 127.0.1.1)

All hosts are connected to a DNS that has authority for domain.tld

# ? Feb 4, 2010 10:04

frumpus: Nov 28, 2005

<-- n00b

I'm installing Debian 5.0.4 from cd iso. I've got disk 1 and the install is going smoothly. Debian's site asks that I not download the other disks unless I know I need packages on them, which makes sense. However, I can not for the life of me locate an index that tells me what packages are available on what disks.

Does anyone know where I can find this?

# ? Feb 4, 2010 18:28

bitprophet: Jul 22, 2004; Taco Defender

frumpus posted:

<-- n00b

I'm installing Debian 5.0.4 from cd iso. I've got disk 1 and the install is going smoothly. Debian's site asks that I not download the other disks unless I know I need packages on them, which makes sense. However, I can not for the life of me locate an index that tells me what packages are available on what disks.

Does anyone know where I can find this?

I honestly don't know why anybody uses the full CDs unless they're doing an offline install; I highly suggest obtaining the netinst version of the installer. This will be a very small CD image which will then just grab the packages you need from the mirrors directly. No worries about popping in Disk 3 of 5 or whatever.

As a minor added bonus, post-install your packages will be fully up to date with e.g. security patches, instead of being whatever was last stuck on the CDs a year or three ago.

# ? Feb 4, 2010 19:45

Bob Morales: Aug 18, 2006; ~~Just wear the fucking mask, Bob~~

I don't care how many people I probably infected with COVID-19 while refusing to wear a mask, my comfort is far more important than the health and safety of everyone around me!

frumpus posted:

<-- n00b

I'm installing Debian 5.0.4 from cd iso. I've got disk 1 and the install is going smoothly. Debian's site asks that I not download the other disks unless I know I need packages on them, which makes sense. However, I can not for the life of me locate an index that tells me what packages are available on what disks.

Does anyone know where I can find this?

You can go based on the jigdo files, but not the ISO files

http://www.debian.org/CD/jigdo-cd/#search

I would just use the first CD, or a net install, and then install the packages you want afterwards. Debian contains something like 3,000+ packages, you'll never use 95% of them, they're just there because they can. I think that's dumb and a waste but oh well.

bitprophet posted:

I honestly don't know why anybody uses the full CDs unless they're doing an offline install; I highly suggest obtaining the netinst version of the installer. This will be a very small CD image which will then just grab the packages you need from the mirrors directly. No worries about popping in Disk 3 of 5 or whatever.

Not everyone has a fast internet connection, or they might be installing from more than one computer, or they might not have a out-of-the-box supported network interface.

# ? Feb 4, 2010 19:57

frumpus: Nov 28, 2005

bitprophet posted:

I honestly don't know why anybody uses the full CDs unless they're doing an offline install; I highly suggest obtaining the netinst version of the installer. This will be a very small CD image which will then just grab the packages you need from the mirrors directly. No worries about popping in Disk 3 of 5 or whatever.

As a minor added bonus, post-install your packages will be fully up to date with e.g. security patches, instead of being whatever was last stuck on the CDs a year or three ago.

Unfortunately a net install is not an option right now.

Bob Morales posted:

You can go based on the jigdo files, but not the ISO files

http://www.debian.org/CD/jigdo-cd/#search

That is helpful thank you. Really hoping I can find a way to browse the contents rather than search them though as there may be a lot of things I don't know I want.

frumpus fucked around with this message at 20:31 on Feb 4, 2010

# ? Feb 4, 2010 20:28

illamint: Jun 26, 2005; According to The Oxford English Dictionary, the word "snapshot" was originally a hunting term.

I've been thinking about it for a while, but I'm wondering, anecdotally, is it worth converting my 4x320GB RAID-5 array from EXT3 to EXT4 or XFS? I'm mostly storing lots of multi-gigabyte files, but there's some smaller stuff in there too.

# ? Feb 4, 2010 21:04

bitprophet: Jul 22, 2004; Taco Defender

Bob Morales posted:

Not everyone has a fast internet connection, or they might be installing from more than one computer, or they might not have a out-of-the-box supported network interface.

Right, which is why I said "unless they're doing an offline install" :v:

but that does sound like the case, so I have nothing further to offer other than sympathy -- the times that I have had the same needs I've also wished for a "I want X Y and Z, which disks do I need?" info dump.

Never found one; doesn't mean one doesn't exist, though.

# ? Feb 4, 2010 21:16

Adbot: ADBOT LOVES YOU

# ? May 10, 2024 17:47

Bob Morales: Aug 18, 2006; ~~Just wear the fucking mask, Bob~~

I don't care how many people I probably infected with COVID-19 while refusing to wear a mask, my comfort is far more important than the health and safety of everyone around me!

I'm sure it's faster and makes way more sense to just apt-get the packages you want. Everything you need is probably on the first 2 CD's, or 1 DVD. No point in downloading a 650MB ISO for 1 or 2 packages.

bitprophet posted:

Right, which is why I said "unless they're doing an offline install" but that does sound like the case, so I have nothing further to offer other than sympathy -- the times that I have had the same needs I've also wished for a "I want X Y and Z, which disks do I need?" info dump.

I really wonder what's on disks 5-31

# ? Feb 4, 2010 21:36

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > The Linux Questions Thread: a bunch of pitfalls, but technically it's possible

«‹›1029 »