My very ownCloud

30 12 2011

A few days ago I was talking with someone about how frustrated I was with the current cloud providers. I mostly just disapprove of how badly my data is treated by the cloud providers. I am not a massive security zealot but I do try to be careful with my data. I have tried several different providers and I just am either not happy with their service or their policy agreements.

While stumbling around on the net in frustration trying to answer this question, I ran into http://owncloud.org/. I was very skeptical at first as it seemed like a very new project. I figured it was worth a try though.

As I prepared for the install (by reading all the docs I could find about the project), I got really super excited. I was going to toss it up on my virtual machine but I got so excited that I ended up doing the install to my ‘production’ system (well about as “production” as it gets for my personal home system…still I don’t like to muck it up). All went well and it was quite easy to get working really fast.

I uploaded a few files and enabled a few plugins and thought ‘Ok. Not perfect. But a decent start.’ As I poked around on the site I really began to like it.

Here is a run down of the plugins I am using.

Files:
Simple enough. You can create text files or directories and you can upload files. When I created a text file I was able to edit it with a basic editor that does most everything I need. I created a few directories and uploaded a few things. The download button worked for files but not for folders. The share button was nice as i could share with a person, group, or public. Most seemed to work as expected.

The only two things I really feel should be addressed are:
1) When uploading a file, all I see is a spinning icon. No display of progress or how much longer the upload will take.
2) It would be nice if I could upload a folder. I tried a few things but they didn’t seem to work. Uploading a file just works, but folders were not working.

Music:
I was surprised and delighted to see that after uploading a (legit!) music MP3 download in the Files section that the tracks appeared in the Music section. The player was working and everything seemed pretty good. The interface was a bit kludgy but it wasn’t hard to figure out. I decided to upload more music. A new album I got for Christmas. An album I ripped to FLAC (like I do all my CD’s). A good 90% of my music collection is in FLAC. The music player stopped working. It didn’t start working until I deleted all the FLAC files.

This was highly disappointing and in my opinon really needs to be addressed…

Contacts:
Simple interface. Handles all the contact information one would expect it to hold.

I do wish there was a notes section.

Calendar:
I set a few appointments/events and while it worked I don’t know how much I will end up using it. I am going to try to use it more because something seems off about it but I can’t place what. Maybe it is just me. But until I can figure out what is bothering me, I can’t write a just and descriptive review of it.

Gallery:
This didn’t seem to auto update after uploading files. Maybe it just had not had the chance to scan. However, a manual scan picked up the files. The preview pictures seemed to be only partial center portions of the image and as such the thumbnail view didn’t really show anything useful in the pics I uploaded. However, clicking on them brought up the full image and I was very please about how easy it was to navigate.

No comments on things to do here.

Bookmarks:
Basic bookmarking feature. Nothing too special. Works just as I expected it to. I found another webpost suggesting that I should be able to link my browser on KDE (Rekonq) to the bookmarks feature. I have not tried this yet, but if it does work that would be very nice as it would be my own bookmark sharing utility.

Over all I am very impressed and I can’t wait to see what they do in the future.

Advertisements




Computer show schwag

15 10 2011

I recently went to a big computer show and I ended up with the normal schwag. Pens, led flashlights, pens with led flashlights, and of course USB flash drives. In fact, I ended up with several flash drives from various vendors. I plugged one in and tried to copy some data to it, only to find that it was TERRIBLY slow. It really was painfully slow. That is why I decided to do some benchmark testing. Nothing too fancy, just the standard utility that ships with Ubuntu LTS 10.04.

System->Administration->Disk Utility

Now because I don’t want to seem too childish in whining about the results, I will not post the vendor names.

As a standard benchmark, a 5$ USB disk from Newegg.
Size 1GB
Max Read 20.1 MB/s
Avg Read 19.2 MB/s
Max Write 5.2 MB/s
Avg Write 4.1 MB/s

A vendor who sells SANs:
Size 1GB
Max Read 20 MB/s
Avg Read 18.6 MB/s
Max Write 5 MB/s
Avg Write 1.9 MB/s

A second vendor who sells SANS:
Size 1GB
Max Read 20.7 MB/s
Avg Read 18.7 MB/s
Max Write 2.8 MB/s
Avg Write 1.6 MB/s

A large enterprise-grade system vendor:
Size 1GB
Max Read 13.1 MB/s
Avg Read 12.3 MB/s
Max Write 3.4 MB/s
Avg Write 3.0 MB/s

A large desktop system vendor:
Size 4GB
Max Read 16.7 MB/s
Avg Read 14 MB/s
Max Write 2.3 MB/s
Avg Write 1.4 MB/s

A Linux vendor:
Size 2GB
Max Read 18 MB/s
Avg Read 16.3 MB/s
Max Write 3.2 MB/s
Avg Write 2 MB/s

A software company for supercomputing:
Size 2GB
Max Read 17 MB/s
Avg Read 12.7 MB/s
Max Write 2.7 MB/s
Avg Write 2.2 MB/s

I found it very interesting (and worthy of this post) simply because of the fact that these flash drives have such terrible write speeds. I understand that each company handed out 100’s of these flash drives and that 5$ each would add up quickly so they went for a much cheaper alternative. Still, it is rather annoying that the write speeds are so terrible. Especially on a 4GB drive. That is a decent amount of space for the files I carry around but at a max of 2.3MB/s transferring files to fill that would be roughly half an hour at peak speeds. Closer to an hour more than likely.

Then again, I really can’t complain too much. I did get a bunch of free flash drives that I am sure I can find a use for.

😀

Thanks to all the vendors who gave me free schwag!





My thoughts on the nine traits of the veteran Unix admin

16 02 2011

I read a interesting article by Paul Venezia called “Nine traits of the veteran Unix admin.” You can find it here. Please, do take a moment to read the article. It might help to understand some of my comments, plus it is a decent read.

Having worked with Linux personally and professionally for over 11 years now along with Unix, HPUX, Irix, and Solaris on occasion, I thought I would offer my thoughts on the article. Especially since he seems to have gotten really close to describing me without ever meeting me…

“No. 1: We don’t use sudo”
The first ‘Wow, that is what I do’ moment was near immediate. The quote “the first thing we do is sudo su -” pretty much describes me to a ‘T’. That is almost always the first command I run when I boot into an Ubuntu live cd.

The only “objection” I have to this is, there are certain times when sudo is necessary. I admin a few servers that are as locked down as tight as we can get them. The root account is locked in every way. The root account and all activity are tracked and monitored. Any admin running commands as root (or su), get flagged on our monitoring system. The only way to work on that box is sudo for the sole purpose of tracking and monitoring what each admin does on that box. It works well and that is about the only time I use sudo.

“No. 2: We use vi”
Damn straight.
Thanks to the Single UNIX Specification, I KNOW that vi is going to be on every system by default. So I KNOW that I have a great editor at my fingertips no matter what system I am on.

“No. 3: We wield regular expressions like weapons”
XKCD jokes aside, if one knows how to wield and document regular expressions one can do near anything.

“No. 4: We’re inherently lazy”
Fully agree. This is why I do everything I can to master the Bash shell. I can automate so much of what needs to be done it isn’t even funny. It may take me a while, and I monitor everything closely, but it gets done and chances are good I don’t have to do it again. Which means I can devote my time to something else AND I have another solution/script in my code repository I can re-use later if need be with little effort.
Oh! Also, screw Angry Birds. Frozen Bubble is where it is at.

“No. 5: We prefer elegant solutions”
I am just going to re-quote Paul on this one because he nailed it perfectly.
“If there are several ways to fix a problem or achieve a goal, we’ll opt to spend more time developing a solution that encompasses the actual problem and preventing future issues than simply whipping out a Band-Aid. This is related to the fact that we loathe revisiting a problem we’ve already marked “solved” in our minds. We figure that if we can eliminate future problems now by thinking a few steps ahead, we’ll have less to do down the road. We’re usually right.”
Can I get an Amen? Amen!

“No. 6: We generally assume the problem is with whomever is asking the question”
I don’t like to be thought of as arrogant, so I try not to be even if others think I have earned it. I will investigate and I will honestly try to find a solution to a one-off-random-happen-chance error, but if I can’t reproduce it chances are I am going to chalk it up to user error. Why? Because it almost always is (sorry; just a bit of the arrogance slipping out). If it happens twice, I will set traps and monitor log files. If I can catch the error, chances are really good that I am going to do what I can to find a solution as long as time, effort, and $$$ permit. I also HATE the Band-Aid philosophy as an admin. I want an elegant permanent solution if I can get one.

“No. 7: We have more in common with medical examiners than doctors”
Once again he hits the nail on the head. Ties right in with the last point. I have detailed monitoring setup with Zabbix for my network and my systems and I expand its capabilities and items it monitors weekly. When something happens, I want the log files to tell me what happened, why, and what I need to do to prevent it from happening again.

“No. 8: We know more about Windows than we’ll ever let on”
Damn monkeys that call themselves Windows Admins. Don’t get me wrong, there are plenty of smart Windows Admins. However, the really bad ones that piss me off are the ones that cause all the trouble. The idiots that really drive me crazy are the ones that wave their MCSE’s around like it is proof they know something, then are completely incapable of thinking outside of the step-by-step guide given to them. These idiots often say ‘It works! Now don’t touch it!’ and won’t do even the basics such as applying security updates (MCSE = Making Computers Susceptible to Exploit). If you ask them to do basic tasks they give stupid excuses (or worse promises) to make you leave so they can hurry back to their desktop games (aka MCSE = Minesweeper Consultant / Solitaire Expert). They have no idea what they are doing. They click this button, then click that button, then click agree, next, next, finish, and then assume that everything is ok. Don’t bother wasting your time telling them they configured something wrong. They will dig up some tutorial book for Windows NT 3.5 as proof for their Windows Server 2008 configuration. These idiots are so scared that someone will force them to learn something new that they are usually the most vicious defenders of buying off the shelf software (God help you if you get caught in their fanatics if someone suggest they learn a new OS! Even another version of MS Windows will set them off!). These idiot admins CAN’T fix problems on their own so they HAVE to buy support and they absolutely depend on management buying whatever they need (MCSE = Management Conned by Something Expensive).
This means whenever something of theirs breaks and that breakage even hints at disturbing anything in the Linux realm, ___I___ have to fix it. Nothing would make me happier then to never touch another Windows box and ditch any and all ‘knowledge’ I have of it, but as long as these idiots are still around I will probably have to fix their servers for them.

PS: Why is it that for every MCSE that I meet who actually knows what he is talking about, I end up dealing with 10 who I am surprised that they managed to fill out their name on the test correctly? Also, if the test is so damn easy and cheap that these idiots can get one, why is it such a highly esteemed certification? That still baffles me to no end….

“No. 9: Rebooting is almost never an option”
I wouldn’t say ‘never an option’…but the last reboots I had were due to kernel updates, a bad nic card I needed to replace, and to run a diagnostic on the root drive that required it to be offline. I will agree with this statement “If the problem occurred once and was “fixed” by a reboot, it’ll happen again.” I have seen this with Windows, OSX, and Linux. Something goes weird and the user reboots every time. Soon they are rebooting several times a day because they never fixed the problem.

My servers have high uptime because they are stable and I know everything running on them and how to fix those programs/services when (if) they have issues. My uptimes can span years and are on legit servers. Not a box that sits in the corner of the room that no one ever logs into.

On the flip side of the coin, I know someone who was hacked pretty bad because he refused to update while he was chasing the holy 5 nines. The hole they got in through had been patched over 9 months previous to the hack. This is why I watch all the security lists for Debian and CentOS (the distros on my servers) and I am more then willing to sacrifice a silly uptime number to apply a security patch.

Lastly, I can’t think of a better way to end this then the statement he had. So I am just going to quote Paul one last time.

“If some of these traits seem antisocial or difficult to understand from a lay perspective, that’s because they are. Where others may see intractable, overly difficult methods, we see enlightenment, born of years of learning, experience, and most of all, logic.”





RHEL 6.0 Beta Review

24 04 2010

I finally got a few minutes to try out the latest Red Hat Enterprise Linux 6.0 beta (downloaded RHEL6.0-20100414.0-AP-i386-DVD1.iso ). Here are my thoughts.

Test rig is nothing special, however, the monitor is capable of going up to 1600×1200 resolution and the vid card is a Intel 910. Yet when the install DVD gets to the installation screen, it just gives me a black screen. I can’t even switch to other terminals. Nothing. I hate it when I can’t see log files or error messages. If decent hardware like this can’t run the install display, how exactly are my servers supposed to handle the install with their 32MB embedded ATI cards attached to the 1280×1024 monitor on the KVM switch?

So I restarted the install but this time I noticed another choice; the ‘Basic Video card’ option. Duh. OK, so that answers my last question. It was slightly annoying at first and I do wonder exactly what kinda video power you have to have to use the first GUI installer screen, but at least they recognized that this option wasn’t going to work for everyone. Kudos for being one step ahead!

The basic video has the old blue and black menu we have seen since the beginning. It asks basic questions like where the install media is located (which most everyone else auto scans for now) and if you want to check the disc for errors and so forth and so on. The usual if you have done Red Hat / CentOS installs before. It will then move you over to a GUI installer for things like root password, time, and location. One thing I really liked was the menu options for partitioning hard drives. Same options as everyone else, but unlike everyone else shoving it into a drop-down option they are all laid out with explanations.

So I chose to use the entire drive, selected the encrypt option just to see what it would do, and review and modify partitions. BAM! Crashed on me. It did give me a nice little error box and the option to save the report. So I clicked save and I got a file menu to the location I want to save it to. I plugged in a USB key and was hoping it would auto mount it. No such luck. So I walked the alt-F[1-6] looking at the terminals. My USB key was recognized and the partitions were recognized; just not mounted. So I mounted it manually and saved out the file. Nice touch on saving the logs. I liked this feature. After saving I got a message to “exit the installer” but clicking it just removed the GUI leaving a black screen and a mouse. The rest of the system continued to work. Shouldn’t it reboot? Ctrl+Alt+Del to the rescue!

Started over again and this time I chose “Create Custom Layout”. I got the familiar disk layout screen and when I selected to delete the old partition…BAM! Crashed again. :-/

Well that’s not good. Maybe it is the disk? Tried Debian Lenny and had no problem formatting the drive into the layout that I wanted. Tried again with Red Hat. No Go (sorry, clicked the wrong button and don’t have this crashlog). Tried again with Red Hat but used the Debian layout I created instead of creating a new one. It worked! Don’t know why Red Hat didn’t like to format my drive.

I was planning on a full desktop, but as soon as I saw “Minimal Installation” I clicked it. I missed that so much from the CentOS 5 installs I did. 228 Packages installed using only 636M!! 😀 This is exactly how I like seeing the start of a stripped-down lean and mean server ready to be built up for any task!

Of course, I really want to mess around with it so the first thing I do is start installing packages…still minimal install is great news for me.

Since I chose a minimal install…I got a minimal install! 😀 This also means I didn’t have network at all nor any of the tools I usually use to work on the network. Ifconfig was still there and all but not much else (as expected!). So I had to vi /etc/sysconfig/network-scripts/ifcfg-eth0 and add “BOOTPROTO=dhcp” so that I could “service network restart” and get an IP address (this statement is more for me when I do this again and can’t remember what I did this time to fix it; it is expected behavior for a bare minimum install).

There were three things I just HAD to check and verify.
1) The Kernel
You have no idea how happy I am to see the “2.6.32-19.el6.i686” kernel! I have a bunch of tests I am itching to try out (starting with PAPI) that were a pain with the old 2.6.18 kernel (looking at you perfctr).
2) Python
Version: 2.6.2 Release: 5.el6
WHOOHOOO!! I have had issues getting Python 2.6 working correctly without breaking things on RHEL/CentOS 5.
3) Gcc and Gnat
Version: 4.4.3 Release: 1.el6

I am so happy right now.

Unfortunately, I am out of time for today and the rest of my testing will have to wait until later…

If anyone has something particular that they want to know or would like for me to test, drop me a note below and I will do what I can to test it out.

[Update Edit 2010-05-02] I finally got some time to work with PAPI. Not as extensive as I want (those tests will come later), however, all that was needed to get PAPI 4.0.0 to work was to extract papi.spec from the tar.gz file and ‘rpmbuild -bb –target=`uname -m` papi.spec &> papi.build.out’ Then rpm -ivh to install the rpms. Tada! That’s it! I am so happy. 😀 Will update later when I run a few more PAPI tests.