Greendrake's questions

According to the AWS doc:

When you stop a running instance ... The instance is migrated to a new underlying host computer when it's started.

Here is also slightly different statement from the same doc:

When you start your instance, it enters the pending state, and in most cases, we move the instance to a new host computer. (Your instance may stay on the same host computer if there are no problems with the host computer.)

So what is the rationale behind the need to move instances to a new host when they are restarted?

The only reason this question has emerged is that I found a question in an AWS Certification Exam sample asking whether migration to new host happens, and currently see no way how one would be able to guess the correct answer.

I have ploop disk image created by dumping an OpenVZ container. There are two files:

791     DiskDescriptor.xml
2.4G    root.hdd

Is there any way to extract the whole root file system from there on a non-OpenVZ machine?

I am trying to take a docker container from one machine and run it on another and encountering this error: "Error response from daemon: No command specified".

Below is a simplified example showing the problem:

docker --version
Docker version 1.10.1, build 9e83765
docker pull ubuntu
docker run --name u1 -dit ubuntu:latest
docker export -o exported u1
docker stop u1
docker rm u1
docker import exported ubuntu:imported
docker run --name u1 -dit ubuntu:imported
docker: Error response from daemon: No command specified.

In that example, we first pull an image (ubuntu) and successfully create/run container u1 from it. Then we export that container to a file (exported), stop/remove the container, import the file into a new image (ubuntu:imported) and try to run a new container from it. It fails.

I am getting "No key available with this passphrase." when trying:

sudo cryptsetup open --type luks /dev/sdc storage --key-file=/path/to/keyfile

The /path/to/keyfile file contains just the passphrase in plain text.

If I enter the same passphrase when asked interactively:

sudo cryptsetup open --type luks /dev/sdc storage

then it works.

Why would --key-file not work in this case? This is Ubuntu 14.04 @ Linux 3.13.0-68.

I am trying to do yum update and getting "MemoryError". The server (OpenVZ box) has 2G RAM and Percona MySQL server running with 1G InnoDB buffer configured. The OS is CentOS 6.5, kernel 2.6.32-openvz-042stab088.4-amd64.

free -m output:

             total       used       free     shared    buffers     cached
Mem:          2048        728       1319          0          0         21
-/+ buffers/cache:        706       1341
Swap:         2048          0       2048

On the host machine, the box memory configuration is:

vzctl set 101 --physpages 0:2G --swappages 0:2G --kmemsize unlimited --privvmpages 2000M:2G --vmguarpages 2000M:2G --oomguarpages 2G --save

When I stop MySQL, yum update works fine. Why would MySQL make yum update fail if there is still more than 1G RAM free?

The full error output is:

yum update
Loaded plugins: fastestmirror, priorities
Loading mirror speeds from cached hostfile
 * base: mirror.mel.bkb.net.au
 * contrib: mirror.mel.bkb.net.au
 * epel: mirror.overthewire.com.au
 * extras: mirror.mel.bkb.net.au
 * updates: mirror.optus.net
Traceback (most recent call last):
  File "/usr/bin/yum", line 29, in <module>
    yummain.user_main(sys.argv[1:], exit_code=True)
  File "/usr/share/yum-cli/yummain.py", line 285, in user_main
    errcode = main(args)
  File "/usr/share/yum-cli/yummain.py", line 136, in main
    result, resultmsgs = base.doCommands()
  File "/usr/share/yum-cli/cli.py", line 434, in doCommands
    self._getTs(needTsRemove)
  File "/usr/lib/python2.6/site-packages/yum/depsolve.py", line 99, in _getTs
    self._getTsInfo(remove_only)
  File "/usr/lib/python2.6/site-packages/yum/depsolve.py", line 110, in _getTsInfo
    pkgSack = self.pkgSack
  File "/usr/lib/python2.6/site-packages/yum/__init__.py", line 887, in <lambda>
    pkgSack = property(fget=lambda self: self._getSacks(),
  File "/usr/lib/python2.6/site-packages/yum/__init__.py", line 683, in _getSacks
    self.plugins.run('exclude')
  File "/usr/lib/python2.6/site-packages/yum/plugins.py", line 184, in run
    func(conduitcls(self, self.base, conf, **kwargs))
  File "/usr/lib/yum-plugins/priorities.py", line 129, in exclude_hook
    repopkgs_archless = _pkglist_to_dict(conduit.getPackages(repo), repo.priority)
  File "/usr/lib/python2.6/site-packages/yum/plugins.py", line 592, in getPackages
    return self._base.pkgSack.returnPackages(arg)
  File "/usr/lib/python2.6/site-packages/yum/packageSack.py", line 499, in returnPackages
    ignore_case=ignore_case)
  File "/usr/lib/python2.6/site-packages/yum/sqlitesack.py", line 1646, in returnPackages
    pkgobjlist = self._buildPkgObjList(repoid, patterns, ignore_case)
  File "/usr/lib/python2.6/site-packages/yum/sqlitesack.py", line 1610, in _buildPkgObjList
    po = self._packageByKeyData(repo, x['pkgKey'], x)
  File "/usr/lib/python2.6/site-packages/yum/sqlitesack.py", line 754, in _packageByKeyData
    po = self.pc(repo, data)
  File "/usr/lib/python2.6/site-packages/yum/sqlitesack.py", line 187, in __init__
    self._loadedfiles = False
MemoryError

Similar question here but not helpful as I have way more RAM.

I am about to setup a database server (MySQL) in OpenVZ container and I've been wondering how many CPUs I should assign to it. I decided to benchmark it. I compared two OS/MySQL distributions and tested how they performed with 1, 2, 3 and 4 CPUs.

The first software configuration was:

• CentOS release 6.5 (Final)
• mysql Ver 14.14 Distrib 5.1.71, for redhat-linux-gnu (x86_64) using readline 5.1

The second:

• Debian GNU/Linux 7 \n \l
• mysql Ver 14.14 Distrib 5.5.31, for debian-linux-gnu (x86_64) using readline 6.2

Both were running on the same kernel - 2.6.32-openvz-042stab083.2-amd64 #1 SMP Fri Nov 8 17:59:25 MSK 2013 x86_64 GNU/Linux.

All software was installed from packages and used out-of-the-box without any custom config tweaking.

Hardware: 6GB RAM, 1-4 CPUs 3.5 GHz.

For benchmarking I used sysbench with the following scenario:

sysbench --test=oltp --oltp-table-size=1000000 --mysql-db=test --mysql-user=root --db-driver=mysql --mysql-password=d prepare
sysbench --test=oltp --oltp-table-size=1000000 --mysql-db=test --mysql-user=root --db-driver=mysql --mysql-password=d --max-time=60 --oltp-read-only=on --max-requests=0 --num-threads=8 run

In both cases the table's engine was InnoDB.

The output I was looking at was the number of transactions per second. The results were quite stable - the error was less than 1%.

The results were nice and expected for CentOS/MySQL5.1, but very strange for Debian/MySQL5.5:

As you may see, MySQL5.5 on Debian fails to properly take advantage of multiple CPUs. While the performance with 2 CPUs is higher than with 1, it is clearly lower than on CentOS/MySQL5.1. Moreover, it goes down when we add more CPUs on top of 2 which is really weird.

Can someone please explain what is going on there? Why on Earth MySQL would perform worse as we add CPUs?

I am aiming to set up full write access to a 9p share for a KVM guest. Both host and guest have the same users/groups with the same IDs. Both host and guest should be able to write to the share using same usernames and I don't want to distinguish whether a file was written by host or guest. The kvm process is running as root — I set user and group to root in /etc/libvirt/qemu.conf.

In the guest definition on the host, the share is defined as follows:

<filesystem type='mount' accessmode='passthrough'>
  <source dir='/mnt/storage/data'/>
  <target dir='data'/>
  <address type='pci' domain='0x0000' bus='0x00' slot='0x06' function='0x0'/>
</filesystem>

In the guest, the share is mounted as:

mount -t 9p -o rw,trans=virtio,version=9p2000.L,msize=262144 data /mnt/data

The problem is that the root user in the guest cannot write to files/folders owned by non-root user on the host. Even more strangely, the guest's root can rename and delete such files. That is, when on the host machine I create a file as a non-root user, I then simply cannot edit it as the guest's root, even though I can rename and delete it!

I also found that where a folder created on the host under non-root user has its permissions set to 777, the guest's root can write to it (i.e. create files in it). This, however, does not apply to files — they still cannot be edited regardless of permissions.

Both host and guest are running Linux server 3.2.0-4-amd64 #1 SMP Debian 3.2.51-1 x86_64 GNU/Linux. In Debian, SELinux is disabled by default and I did not enable it. I tried all three available 9p access modes (passthrough, mapped and squash) — no difference.

Just wondering if there is anything I can tweak to get it work, or is it just a bug?

Note there is a similar issue reported here: Read/write access for passthrough (9p) filesystems with libvirt/qemu? but unlike that case, I have 100% write access where files owned by root, it's just non-root user files that I cannot write to even being root on the guest.

I am load-testing a web application and studying the impact of some config tweaks (related to disk i/o) on the overall app performance, i.e. the amount of users that can be handled simultaneously.

But the problem is that I hit 100% CPU before I can see any effect of the disk-related config settings. I am therefore wondering if there is a way I could deliberately limit the disk performance so that it becomes the bottleneck and the tweaks I am trying to play with actually start impacting performance. Should I just make the hard disk busy with something else? What would serve the best for this purpose?

More details (probably irrelevant, but anyway): PHP/Magento/Apache, studying the impact of apc.stat. Setting it to 0 makes APC not checking PHP scripts for modification which should increase performance where disk is the bottleneck. Using JMeter for benchmarking.

I am running CentOS 6.4 with OpenVZ on my laptop. In order to provide Internet access for the VEs I have to apply the following rule on the laptop:

iptables -t nat -A POSTROUTING -j SNAT --to-source <LAPTOP_IP>

It works fine.

However, I have to work in different places - office, home, partner's office etc. The IP of my laptop is different in those places, so have to alter the rule above each time I change place.

I have created a workaround which basically determines the IP and applies the rule:

#!/bin/bash
IP=$(ifconfig | awk -F':' '/inet addr/&&!/127.0.0.1/{split($2,_," ");print _[1]}')
iptables -t nat -A POSTROUTING -j SNAT --to-source $IP

The workaround above works. I only still have to execute it manually. Perhaps I could make it a hook executing whenever my laptop obtains an IP address from DHCP - how can I do that?

Also, I am just wondering if there is an elegant way of getting it done in the first place - iptables? Maybe there is a syntax allowing to specify "current hardware ip addres" in the rule?

I am sending email messages from a PHP script to an Exim 4.71 server via SMTP. The server advertises pipelining so I am trying to take advantage of it. However, when I read back the return codes I finally get 554 SMTP synchronization error. On the server side the error reads:

SMTP protocol synchronization error (next input sent too soon: pipelining was advertised): rejected "DATA" H=(localhost) [111.111.111.111] next input="To: [email protected]\nDate: Wed, 12 Jun 2013 11:35:50 +1200\nReturn-Path: [email protected]\nSubject: Blablablablabla"

I am wondering why is it saying "next input sent too soon". I am using pipelining and so should be allowed to send my data not waiting for anything.

Note that when I do not use pipelining it all goes fine.