|
Martytoof posted:A bit of a branch off, but has anyone jumped into AIX from a mainly Linux background? What did you find was most difficult during the adaptation? Was there anything about AIX that exceeded your expectations? I'm looking to expand my horizons beyond Linux to AIX for .. reasons, and I'm debating picking up a little lab Power5 to play with. Just curious what others' experiences were. Yes, though not by choice, and I didn't deal with a lot of the AIX-specific virtualization stuff, which seems to be one of its main selling points. Most of the work figuring out the operating requirements and processes for our software had already been done for me also. My experience was that the userland tools were generally much worse and more annoying to work with, since IBM has focused on keeping compatibility with old Unix apps, and as such everything still works like it did in the 80s. ksh in vi mode is kind of a dick even if you're used to vi. That said, it's not too different (compared to say, VMS) from Linux, and if you're capable of figuring out the differences between different Linux distributions or Linux and BSD you can probably do the same for AIX without much trouble if you're not getting deep into the differences in kernel architecture or other low-level stuff, which I haven't worked with but is probably very different. IBM has excellent docs too, which makes things much easier.
|
# ? Feb 10, 2015 02:50 |
|
|
# ? May 10, 2024 01:01 |
|
annapacketstormaya posted:. AIX is a dead skill. Don't bother. This isn't a "Linux fanboy" argument. I really liked Solaris and AIX 10 years ago, but AIX has nothing substantiative to offer now. Some of the PowerVM stuff is neat, but see above. reading posted:I can't change the new home folder with chown to be my username:username (uid and gid) because it isn't a valid choice. Looking at /etc/passwd, my username isn't included in the list, which I'm sure is because I'm using a liveboot cd. Right now it's root:root which I assume will screw things up when I try to reboot and login as myself. Yes, you can . chowning to root is stupid. ls -ln /yourhomedir Look at the uid and gid. chown uid:gid... I could chown everything to 57341 if I wanted to. It doesn't have to be in passed to work, or NFS administrators would have a huge mess on their hands Thermopyle posted:edit: I guess this is more of a vmware player issue than a linux issue since if I switch the vmware network adapter to NAT it works, but as soon as I switch it back to bridged it stops working... Bridging not in promiscuous? Windows firewall on and blocking it?
|
# ? Feb 10, 2015 15:30 |
|
Why is AIX going away?
|
# ? Feb 10, 2015 15:48 |
|
Tab8715 posted:Why is AIX going away? Cost, mostly. I've seen at least two large companies begin dumping AIX in favor of Linux for enterprise applications, mostly because the pool of talent for developers is larger and Linux is much, much cheaper to support than AIX (especially on POWER hardware). Support contracts with IBM are stupidly expensive.
|
# ? Feb 10, 2015 15:53 |
|
Tab8715 posted:Why is AIX going away? Not going away, just a dead skill. Why learn something with diminishing market share and a producer more interested in selling consulting (on any OS, not just AIX? PowerVM is a different story
|
# ? Feb 10, 2015 16:24 |
|
The reason I brought up AIX was that I could possibly be put in a position to touch some AIX assets in the near future so I was hoping to get a bit of a leg up. Dying skillset perhaps, but I'm also the guy teaching himself z/OS so I'm picking the dying-est skillsets to bone up on
|
# ? Feb 10, 2015 17:41 |
|
Martytoof posted:The reason I brought up AIX was that I could possibly be put in a position to touch some AIX assets in the near future so I was hoping to get a bit of a leg up. Dying skillset perhaps, but I'm also the guy teaching himself z/OS so I'm picking the dying-est skillsets to bone up on It's a commercial UNIX that actually has sane packaging. Get used to sysv coreutils. Use smit. Love smit. (smit will actually show you what commands it ran in its log, which can be invaluable for learning odd AIX bits). There are some odd commands for disk management et al. Google's your friend if you need these things
|
# ? Feb 10, 2015 17:46 |
|
reading posted:I can't change the new home folder with chown to be my username:username (uid and gid) because it isn't a valid choice. Looking at /etc/passwd, my username isn't included in the list, which I'm sure is because I'm using a liveboot cd. Right now it's root:root which I assume will screw things up when I try to reboot and login as myself. Don't use the username or group name, use the numerical identifiers instead. example, assuming your user number is 1001 and the group number for your user is 501(common defaults for an only user) code:
|
# ? Feb 10, 2015 17:46 |
|
RFC2324 posted:Don't use the username or group name, use the numerical identifiers instead. I was able to log in by deleting the old .Xauthority file in my home dir, and once I logged in as myself I chowned everything back to username:username. Thanks for the help everyone.
|
# ? Feb 10, 2015 18:19 |
|
Cidrick posted:Cost, mostly. I've seen at least two large companies begin dumping AIX in favor of Linux for enterprise applications, mostly because the pool of talent for developers is larger and Linux is much, much cheaper to support than AIX (especially on POWER hardware). Support contracts with IBM are stupidly expensive. I guarantee you there will be at least a few accounts that cling onto their AIX installs until the end of time and continue paying out the rear end for extended life support contracts because they can never make changes to production ever. There are places that still have their IBM mainframes runnning, so why not AIX?
|
# ? Feb 10, 2015 22:41 |
|
annapacketstormaya posted:I guarantee you there will be at least a few accounts that cling onto their AIX installs until the end of time and continue paying out the rear end for extended life support contracts because they can never make changes to production ever. There are places that still have their IBM mainframes runnning, so why not AIX? AIX is still pretty prevalent in healthcare IT, IIRC.
|
# ? Feb 10, 2015 23:04 |
|
annapacketstormaya posted:I guarantee you there will be at least a few accounts that cling onto their AIX installs until the end of time and continue paying out the rear end for extended life support contracts because they can never make changes to production ever. There are places that still have their IBM mainframes runnning, so why not AIX? Sure, I don't disagree. But my point is that the talent pool for AIX is shrinking. How many existing enterprises have you heard of moving *to* AIX, or new shops deciding "Let's run AIX instead of Linux"? Every instance of it I've seen in the wild has been running on AIX for a very long time, and they're comfortable with it and like the stability of it, and that's their call. For what it's worth, I think AIX is a very stable and reliable OS. I just don't think it's a platform worth learning unless you have a pretty good reason, like - oh poo poo, we merged with another company and firing all their AIX guys, and now I have to support it.
|
# ? Feb 11, 2015 01:05 |
|
captkirk posted:AIX is still pretty prevalent in healthcare IT, IIRC. I was in an AIX workshop once where the instructor bragged about how the FAA (or at least part of it) was still running AIX 3.1 because it was rock solid and never went down. At first I was baffled as to who at the FAA hadn't retired that could support a (then) 22-year-old operating system, but then I realized that IBM was supporting it all so they just brought in whomever they needed.
|
# ? Feb 11, 2015 01:12 |
|
Cidrick posted:Sure, I don't disagree. But my point is that the talent pool for AIX is shrinking. How many existing enterprises have you heard of moving *to* AIX, or new shops deciding "Let's run AIX instead of Linux"? Every instance of it I've seen in the wild has been running on AIX for a very long time, and they're comfortable with it and like the stability of it, and that's their call. The corollary, of course, is that it's useless unless you plan on running that specific software.
|
# ? Feb 11, 2015 01:53 |
|
captkirk posted:AIX is still pretty prevalent in healthcare IT, IIRC. It is. Guess why I worked with it? Cidrick posted:Sure, I don't disagree. But my point is that the talent pool for AIX is shrinking. How many existing enterprises have you heard of moving *to* AIX, or new shops deciding "Let's run AIX instead of Linux"? Every instance of it I've seen in the wild has been running on AIX for a very long time, and they're comfortable with it and like the stability of it, and that's their call. This is also why I ran away screaming from it as soon as I could. More because they were using AIX as a platform to run Intersystems Cache, but still. To their credit, they were making a fairly large push to try and move forward to VMWare/RedHat as their chosen platform.
|
# ? Feb 11, 2015 01:54 |
Bumping this one again...is this how it's normally done?fletcher posted:I need multiple services to have access to the same SSL key, should I create a new group called 'ssl' or something for that and add the service users to it?
|
|
# ? Feb 12, 2015 01:59 |
|
Cidrick posted:Sure, I don't disagree. But my point is that the talent pool for AIX is shrinking. I think this is as good a reason to at least be familiar with it as any. With IBM moving to embrace z/Linux on big iron I think the writing is on the wall, but it can never hurt to become familiar with the tool. I may actually end up working on it so I have a somewhat legit reason to look into it, but if I were a jack of all trades UNIX guy (and I was) then it would be tempting just to put that notch in my bedpost
|
# ? Feb 12, 2015 02:17 |
|
I know someone who posts here occasionally that sometimes takes a run at puppetizing AIX. It usually doesn't go very well.
|
# ? Feb 12, 2015 05:31 |
|
fletcher posted:I need multiple services to have access to the same SSL key, should I create a new group called 'ssl' or something for that and add the service users to it? That is a valid and very sensible way to do it. I would expect most/all service processes to read their keys and other configuration data before any security precautions like shedding any "extra" group memberships, so if that solution does not work with some service, I think it should be considered a minor bug in the service. Of course, you might want to make some sort of a note somewhere, saying something like "this SSL certificate/key is used by services X, Y and Z. When it is replaced, remember to test them all afterwards."
|
# ? Feb 12, 2015 10:05 |
|
Misogynist posted:For whatever it's worth, we found Node 0.10.33 to be an extremely problematic release. You'll probably have much better luck with 0.10.35. But we'll probably be better at helping you with your specific problem if you let us know what that problem is, rather than making us guess. Solved but I was getting this error: code:
I'm still not totally clear on what's happening when building node fibers/installing packages but that's where my problem was. In this case, bcrypt was throwing errors when I ran a meteor application with node bundle/main.js. I was building on homebrew and deploying to Debian Wheezy, and the solution was to manually install bcrypt and copy it into bundle/programs/server as it looks like the binary isn't cross compatible (thanks stackoverflow). It happened with a few other packages when I deployed TelescopeJS so I just started building the project tarballs on Linux and I've been fine. Frustrating though
|
# ? Feb 12, 2015 17:13 |
|
Roargasm posted:Solved but I was getting this error: Are you saying that OSX and Linux aren't binary compatible? Why is this a surprise? Build environment should match production. OSX is not a substitute for Linux. Homebrew is general crap
|
# ? Feb 12, 2015 17:54 |
|
I'm trying to get logstash working on Fedora 21 and I'm having no luck at all. Following this guide: http://logstash.net/docs/1.4.2/tutorials/getting-started-with-logstash When I do java -version I get openjdk version "1.8.0_31" I was able to download the file with curl and unpack it. When I tried to run it with the -e 'input.... All I get is a blank prompt that doesn't echo like I expected. How do I even start troubleshooting? Am I out of luck until I have a better understanding of how Linux actually works? Edit: I'M AN IDIOT So, it turns out that the problem is that logstash takes a minute to get started. I'd been typing some test commands, getting nothing and then using Ctrl-C While making this post, all the output I expected popped up. Dr. Arbitrary fucked around with this message at 19:28 on Feb 12, 2015 |
# ? Feb 12, 2015 19:26 |
|
Dr. Arbitrary posted:I'm trying to get logstash working on Fedora 21 and I'm having no luck at all. Logstash can take a while to start up, have you tried waiting 60s , then typing stuff at the command line to see if it echos json back at you ? edit: yep.
|
# ? Feb 12, 2015 19:30 |
|
Dr. Arbitrary posted:I'm trying to get logstash working on Fedora 21 and I'm having no luck at all.
|
# ? Feb 12, 2015 21:09 |
|
Misogynist posted:In general, you should assume that any Java service is going to take about 60 seconds to start doing anything at all. But how long do you wait for it to start working well?
|
# ? Feb 12, 2015 22:44 |
|
RFC2324 posted:But how long do you wait for it to start working well? 20 years and counting.
|
# ? Feb 12, 2015 23:10 |
|
I've got a bash script that runs some php scripts (that I didn't write), sleeps for 5 minutes and then loops and runs them again, forever. These php scripts are poo poo from someone else, but I don't want to spend ages rewriting them. By "these php scripts are poo poo" I mean that for whatever reason they freeze up a couple times a day which means I've got to keep checking on them. Is there an easy way to check for any console output and if it hasn't changed in awhile kill my little bash script and run it again? I mean, I could rewrite it in Python and I could do it there easy as pie, but I'm lazy and thought I'd ask if there was some cool linux way of doing this...
|
# ? Feb 12, 2015 23:18 |
|
Martytoof posted:I think this is as good a reason to at least be familiar with it as any. With IBM moving to embrace z/Linux on big iron I think the writing is on the wall, but it can never hurt to become familiar with the tool. I may actually end up working on it so I have a somewhat legit reason to look into it, but if I were a jack of all trades UNIX guy (and I was) then it would be tempting just to put that notch in my bedpost If you can find one, an older version of the O'Reilly Essential System Administration by AEleen Frisch dealt with AIX and even IRIX, I think it was 2nd Edition, but 3rd edition still has AIX.
|
# ? Feb 13, 2015 05:21 |
|
Thermopyle posted:I've got a bash script that runs some php scripts (that I didn't write), sleeps for 5 minutes and then loops and runs them again, forever. These php scripts are poo poo from someone else, but I don't want to spend ages rewriting them. By "these php scripts are poo poo" I mean that for whatever reason they freeze up a couple times a day which means I've got to keep checking on them. A quick and dirty way would be to redirect the script's output to a file. Then every so often (via a cron job) have a simple "caretaker" script look at the size of that file and store it. If the size of the output file hasn't changed by the next time the caretaker is run, restart your original script. Powered Descent fucked around with this message at 07:07 on Feb 13, 2015 |
# ? Feb 13, 2015 07:04 |
|
If you're on a system with systemd, use the WatchdogSec feature in conjunction with a watchdog service. From the systemd docs: quote:WatchdogSec= WatchdogSec=<some # of seconds> ExecStart=/bin/my/janky_script.php | tee /mylogfile | watchdog Restart=always Where watchdog is a little program or script you'll have to write that waits for input on stdin and calls sd_notify(0, "WATCHDOG=1") when it receives anything. If nothing arrives within <some # of seconds>, systemd will assume the script is dead and restart it.
|
# ? Feb 13, 2015 07:36 |
|
Thermopyle posted:I've got a bash script that runs some php scripts (that I didn't write), sleeps for 5 minutes and then loops and runs them again, forever. These php scripts are poo poo from someone else, but I don't want to spend ages rewriting them. By "these php scripts are poo poo" I mean that for whatever reason they freeze up a couple times a day which means I've got to keep checking on them. There is a little "timeout" utility in coreutils that runs the program you pass to it, and sends it a SIGTERM after a specified time.
|
# ? Feb 13, 2015 15:49 |
|
I'm trying to install the Gnome desktop on Fedora 21 server using VMware workstation. It hangs on starting Gnome. I was able to figure out how to get into the console and was able to type: sudo systemctl status gdm I don't see anything useful looking other than "Failed to start GNOME Display Manager" On a possibly related note, I haven't been able to figure out how to install VMware tools. Any ideas on how I can start fighting this problem?
|
# ? Feb 13, 2015 17:16 |
|
Dr. Arbitrary posted:I'm trying to install the Gnome desktop on Fedora 21 server using VMware workstation. I guess my first question is: Is there any reason you are using Fedora 21 Server instead of Fedora 21 workstation?
|
# ? Feb 13, 2015 18:27 |
|
Dr. Arbitrary posted:I'm trying to install the Gnome desktop on Fedora 21 server using VMware workstation. It's probably lack of 3d accel. Tell it to install the VMware tools. Then type "mount" and look for the CDROM. If it isn't mounted, mount it somewhere "mount /dev/cdrom /some/path" yum -y groupinstall (or dnf -y groupinstall) "Development Tools" yum -y install perl You need kernel-header, kernel-devel, gcc, and a few others to build the extensions. cd /path/to/cd/mountpoint Find the "VMwareTools" tarball (I think this is still there) and move it to /tmp. Extract it and ./vmware-install.pl It'll build and install the kernel modules. Alternatively, "yum -y install open-vm-tools", which does much of the same stuff, I think.
|
# ? Feb 13, 2015 18:31 |
|
ratbert90 posted:I guess my first question is: Is there any reason you are using Fedora 21 Server instead of Fedora 21 workstation? This is all in a lab environment so I'm intentionally doing things in ways that are not convenient. My current "goal" is to set up an ELK stack that does anything, even if it's just counting the letter frequency distribution of my log files.
|
# ? Feb 13, 2015 18:44 |
|
Dr. Arbitrary posted:It hangs on starting Gnome. Looks like the GNOME Display Manager is installed but not working, or it gives up after failing to start the X server (=the cornerstone of your GUI environment). Look into the log files. In particular, /var/log/gdm.log and/or /var/log/Xorg.0.log should be helpful.
|
# ? Feb 13, 2015 18:46 |
|
evol262 posted:Extract it and ./vmware-install.pl I'm getting stuck here. It says sudo: unable to execute ./vmware-install.pl: Permission denied
|
# ? Feb 13, 2015 18:49 |
|
Dr. Arbitrary posted:I'm getting stuck here. It says You have insufficient permission to install, type sudo <scriptname>
|
# ? Feb 13, 2015 18:54 |
|
Dr. Arbitrary posted:I'm getting stuck here. It says Is the script located somewhere mounted noexec?
|
# ? Feb 13, 2015 18:56 |
|
|
# ? May 10, 2024 01:01 |
|
Dr. Arbitrary posted:I'm getting stuck here. It says "sudo perl vmware-install.pl" It's probably not chmod +x
|
# ? Feb 13, 2015 19:00 |