Peter's questions

I'm running Docker 3.6.0 on a Mac. I only run one container, which is a postgres 9.6 instance. Over the last few months, I've noticed the stability of the postgres instance get worse and worse and eventually I just restart the container multiple times a day and then it keeps crashing later on.

If I stop Docker completely, then start it again, I run "container logs " (without even starting any container) and it very rapidly logs this line repeatedly:

FATAL:  password authentication failed for user "postgres"
DETAIL:  Password does not match for user "postgres".
    Connection matched pg_hba.conf line 95: "host all all all md5"

I've tried to stop any programs that would be making a connection to this database (that I know of) and yet I still keep seeing this logged in fast succession.

I'm not totally sure if this logging is related to my problem with stability. Ultimately, I would like for the postgres instance to not crash all the time.

Docker is running with port 5432 mapped from my Mac to the docker instance (same port).

If I look within the container at: "/var/lib/postgresql/data/pg_hba.conf", I see it ends with:

host all all all md5

I'd like to know what happens to connections when nfs is failed over to another server (drbd). If I have a file open in vim, it will lock up I think only sometimes come back.. after a long while?

What are some common ways of knowing whether it blocks on io, kills connections etc when the fail over occurs?

So this is using a floating IP between 2 servers using nfs with drbd volume.

So I have a particular ext4 drive that doesn't allow me to execute files. I have the following example in bash where even with sudo and with the file having execute permissions for all users, the file does not execute. If I move the file to an alternate partition, it works fine. Thoughts?

$ ./pycharm.sh
bash: ./pycharm.sh: Permission denied
$ sudo ./pycharm.sh
sudo: unable to execute ./pycharm.sh: Permission denied
$ sudo -s
# ./pycharm.sh
bash: ./pycharm.sh: Permission denied
# lsattr pycharm.sh
-------------e-- pycharm.sh
# ls -la pycharm.sh 
-rwxr-xr-x 1 me me 5908 Nov 20 03:18 pycharm.sh

I'm trying to configure static IPs on 2 vagrant boxes so they can communicate with each other. With the following config, the "db" node comes up with an eth1 and eth2 with 192.168.50.4 and 192.168.50.5 respectively. The second "web" node, correctly brings up only an eth1 with 192.168.50.5. Why does the first node have both addresses?

Vagrant.configure("2") do |config|
  config.vm.provision "shell", inline: "echo Hello"

  config.vm.define "web" do |web|
    web.vm.box = "ubuntu/trusty64"
    config.vm.network "private_network", ip: "192.168.50.5"
  end

  config.vm.define "db" do |db|
    db.vm.box = "ubuntu/trusty64"
    config.vm.network "private_network", ip: "192.168.50.4"
  end
end

Anyone know of a way to ascertain the last accessed date for s3 files/buckets or if this is even recorded? Doesn't seem visible via the GUI and a basic try of s3cmd and boto shows last modified date but not accessed.

I need to access data from a server that is behind a proxy. Thus, I can normally ssh to the proxy and then ssh to the destination to perform administration. However, not sure how to achieve this in bash. I do have a local ssh alias that will tunnel through the proxy automatically when i "ssh destination", but this must be in my environment or something and not used by the script.

In essence, is it possible to "ssh proxy ssh destination $commmand" in bash? I have tried the this exact command and had no success.

What is a clean way to install Flash on ubuntu server 12.10? The package flashplugin-installer doesn't seem to be available any more. Is it no longer packaged? Or would I need to do something like step 3 in https://askubuntu.com/questions/11/how-do-i-install-adobe-flash-player even though that seems to be for Ubuntu Desktop edition, not Ubuntu Server.

This is for incorporation with Firefox for testing.

I have a Debian Wheezy install that is currently showing 100% disk usage, yet the files in / do not seem to add up to the sum of the disk as reported in a df -h output (as tested with a du -sh /). I have since rebooted, so this would have cleared out any open files that could have been deleted but still remained open.

Output of mount:

/dev/disk/by-uuid/blahblah-3431-43i9-ac71-blahlah on / type ext4 (rw,relatime,errors=remount-ro,user_xattr,barrier=1,data=ordered)

Output of df -h:

rootfs                                                   36G   33G  810M  98% /
/dev/disk/by-uuid/blahblah-3431-43i9-ac71-blahlah   36G   33G  825M  98% /

Output of du -sh of / (excluding media, as that dir has directories mounted from other drives):

# du -sh --exclude=media *
7.4M    bin
18M     boot
0       dev
11M     etc
13G     home
0       initrd.img
116M    lib
4.0K    lib64
16K     lost+found
4.0K    mnt
341M    opt
du: cannot access `proc/5744/task/5744/fd/4': No such file or directory
du: cannot access `proc/5744/task/5744/fdinfo/4': No such file or directory
du: cannot access `proc/5744/fd/4': No such file or directory
du: cannot access `proc/5744/fdinfo/4': No such file or directory
0   proc
134M    root
1.1M    run
7.6M    sbin
4.0K    selinux
4.0K    srv
0       sys
56K     tmp
6.4G    usr
2.4G    var
0       vmlinuz

I'm having a problem where adding an opennms scheduled outage is not functioning. Considering I can find limited documentation about how to do this through their interface (I suppose it's supposed to intuitive), I'd like to confirm these are the correct steps.

I begin by clicking on a node and clicking to "schedule outage" (Home / Admin / Scheduled Outages / Edit). I click a node and then on the main form page, I; search for a node and click "Add" (with or without "Add with path"), set "outage type" to "specific" and click "set", set a time between 2014 00:00:00 and 2014 23:59:59 (to set outage for all day) and click "add outage", then under "applies to" I select "all notifications" and scroll to the bottom and click "Save".

This will then redirect me to the scheduled outage log, of which I can see some entries but my entry that was just created does not appear and a scheduled outage is not set. Before I start looking into any potential server problems, can somebody please let me know if these steps are correct and not just user-error causing my problem?

I'm finding this very tricky in a multi node environment, with code under revision control. AKA. multiple frontends and single database. I have a deployment process that pushes a git repo to the servers, but obviously if I update Wordpress from within the admin panel, it will update the files to one FE. Then I would need to copy over the new files to the other FE nodes. Plus, whenever these changes are written when Wordpress updates on a node, it writes code into the git repo. As such, it then breaks the auto deploys that perform 'git pulls', as it then has untracked changes and refuses to pull in new deploys unless manually intervened.

How does one easily keep Wordpress updated in a multi node (load balanced) environment?

What are my options if the version of PHP packaged with Debian Wheezy is too old and I want to run something newer? Besides switching to Ubuntu stable or running Debian Jessie (not great for production), is there a way to only install the new version of php and continue to receive package updates for it?

It's possible to run "ssh user@server $command". This command does not write to bash history, as no shell is opened. I am wondering what signs are present on a server to show evidence of this command. I suppose at the time it is executing, it will be present in a ps output and maybe if you run "w" to see who is connected and what they are running? Otherwise, if you don't see the command whilst it is executing, is there any other information logged on the server about this? This is assuming you are running a command that isn't modifying files or somehow leaving traces of changes being committed.

Is there a unix utility that offers logging of such commands?

Just say you are terminating http SSL connections on an intermediary server like Cloudflare, this will decrypt the data and then forward it onto the server for processing. What is the performance gain of doing this? If you unencrypt the data at the intermediary and then send it to the server, doesn't this mean the data can be sniffed in between? And if you get around this by making a further SSL connection from intermediary to server, doesn't this create so much overhead that there is no performance benefit to terminating at an earlier hop in the first place?

I am having issues with a Tomcat service that crashes (stops responding to requests) and needs to be sigkill'd and restarted every few days. My graphing shows that HEAP memory usage gets to about 3gb during the peaks and this is the most obvious symptom of the issue. From what I have read, adding in Xmx max memory limit for jvm is what dictates how much HEAP can be allocated. Xmx is set to over 10gb.

I have also played around with the argument MaxPermSize to no avail. I think this setting defines memory allocation for some application memory usage, but afaik it is different to HEAP memory itself.

Now I am wondering if there is an alternate setting I can pass in that defines the /actual/ HEAP memory limit? Otherwise, any tips on investigating the problem would be good, as I can't positively say it is memory limits being hit and the error logs do not show much.

Just say you have 50 vhost files configured and your server is under load for apache, what is a good way to see which site is getting trafficked the most? Obviously tailing access logs can only be done one at a time, you could use apachetop, but that only points to one file at a time (I think). You could check the size of the access logs on the filesystem and the largest one is probably getting the most hits. Any better ideas to achieve this?

I have a domain that once you perform a dig mx on it, it returns a set of name servers in the authority section that look like the list below. And yet, when I perform a whois check on it, it returns a completely different set of nameserver records. What are the possible circumstances that could lead to this discrepancy? So far all I can think of is a recent redelegation in name servers had occurred (and the change hasn't been fully propagated yet) or or the domain had expired or been transferred between registrars, and neither of these things seem to have occurred.

;; AUTHORITY SECTION: com.au. 78300 IN NS w.au. com.au. 78300 IN NS x.au. com.au. 78300 IN NS y.au. com.au. 78300 IN NS z.au.

I am looking for some clarification on GET. If I have a URI such as https://google.com/something/ and I try to access that webpage, web servers interpret this as a GET request for /something/ rather than a GET for https://google.com/something/? Server access logs will commonly only log the /something/ part when visitors access the site. Yet, if I perform a curl on that full address, the internet tells me that curl is preforming a GET on the full URI. As such... what is the distinction here?

Also, what exactly does the shortened address, /something/, represent? Is it a file that resides in the root of the website being accessed? What if I set up a rewrite for /something/ to 301 to a particular address and /something/ is not actually a file or directory; how is a request like that processed?

And lastly, is https://username:[email protected]/ still the standard for accessing a website over https with http auth?

So if I open up a Mysql CLI and run a select statement that includes many columns and row content that is long, mysql will render this depending on my window resolution. If my screen is small and a row doesn't fit, it will cut in half and display the second half of the row on a second line. I am looking for a way to output the rows as a continuous line (so they can be read easily) and then use a command to tab to the right and see the second half of the content. Any ideas on how to achieve this?

I'm trying to troubleshoot a database import problem and want to duplicate the environment onto another server. This will require installing an older version of mysql, but the packages that are listed are only showing a recent version. I'm currently running debian wheezy 7.1 and what was installed was the packaged 5.5.31. What is the official way to install an older copy? I guess I could hunt around Google and hope to find some files of the same version to install from source, but this doesn't seem like a reliable method.

Let's say I am getting no connection to a website over http, ie. "Oops! Google Chrome could not connect to x" returns in a browser. What steps can I take to determine if the server is completely offline or whether there are still some services active?

So far, I have:

• confirmed an A record is returning.

• getting 'destination host unreachable' from ping.

• nmap replies with no services, saying that 1000 scanned ports are filtered.

• nping fails to initiate a TCP handshake saying 'no route to host'.