bwizzy's questions

I'm setting up a MariaDB Galera cluster and I can’t seem to get the nodes to join each other. They both start up without errors but never seem to join, the status is always disconnected:

MariaDB [(none)]> show status like 'wsrep%';
+--------------------------+----------------------+
| Variable_name            | Value                |
+--------------------------+----------------------+
| wsrep_cluster_conf_id    | 18446744073709551615 |
| wsrep_cluster_size       | 0                    |
| wsrep_cluster_state_uuid |                      |
| wsrep_cluster_status     | Disconnected         |
| wsrep_connected          | OFF                  |
| wsrep_local_index        | 18446744073709551615 |
| wsrep_provider_name      |                      |
| wsrep_provider_vendor    |                      |
| wsrep_provider_version   |                      |
| wsrep_ready              | ON                   |
+--------------------------+----------------------+

first node is started with: "sudo service mysql start —wsrep-new-cluster" (also tried passing in the cluster address just as —wsrep_cluster_address=gcomm://) as well as putting the empty address list in my.cnf

second node is started with: "sudo service mysql start" with cluster address in my.cnf as gcomm://172.16.56.130,172.16.56.131

no matter what the cluster status is always “Disconnected” and cluster size is 0.

I had some settings for SST to use xtrabackup, iptables firewall was on, etc but in order to try and find the issue I turned everything off and went with the most simple config & still the same result.

Here's the my.cnf

# MariaDB database server configuration file.

[client]
port        = 3306
socket      = /var/run/mysqld/mysqld.sock


[mysqld_safe]
socket      = /var/run/mysqld/mysqld.sock
nice        = 0

[mysqld]
#
# * Basic Settings
#
user        = mysql
pid-file    = /var/run/mysqld/mysqld.pid
socket      = /var/run/mysqld/mysqld.sock
port        = 3306
basedir     = /usr
datadir     = /var/lib/mysql
tmpdir      = /tmp
lc_messages_dir = /usr/share/mysql
lc_messages = en_US
skip-external-locking
bind-address        = 0.0.0.0

# * Fine Tuning
max_connections     = 100
connect_timeout     = 5
wait_timeout        = 600
max_allowed_packet  = 16M
thread_cache_size       = 128
sort_buffer_size    = 4M
bulk_insert_buffer_size = 16M
tmp_table_size      = 32M
max_heap_table_size = 32M


# * MyISAM
myisam_recover          = BACKUP
key_buffer_size     = 128M
table_open_cache    = 400
myisam_sort_buffer_size = 512M
concurrent_insert   = 2
read_buffer_size    = 2M
read_rnd_buffer_size    = 1M

# * Query Cache Configuration
query_cache_limit       = 128K
query_cache_size        = 64M
log_warnings        = 2
slow_query_log_file = /var/log/mysql/mariadb-slow.log
long_query_time = 10
log_slow_verbosity  = query_plan
log_bin         = /var/log/mysql/mariadb-bin
log_bin_index       = /var/log/mysql/mariadb-bin.index
expire_logs_days    = 10
max_binlog_size         = 100M

# * InnoDB
innodb_buffer_pool_size = 1G
#innodb_log_file_size = 100M
innodb_log_buffer_size  = 8M
innodb_file_per_table   = 1
innodb_open_files   = 400
innodb_io_capacity  = 400
innodb_flush_method = O_DIRECT


[mysqldump]
quick
quote-names
max_allowed_packet  = 16M

[mysql]

[isamchk]
key_buffer      = 16M


# MariaDB / Galera Cluster Settings
wsrep_provider         = /usr/lib/galera/libgalera_smm.so
wsrep_provider_options = "gcache.size=32G"
wsrep_cluster_address  = "gcomm://172.16.56.130,172.16.56.131"
wsrep_cluster_name     = 'my_galera_cluster'
wsrep_node_address     = '172.16.56.131'
wsrep_node_name        = 'lou-dev-sc-db2'
wsrep_sst_method       = xtrabackup
wsrep_sst_auth         = root:rootPa$$
wsrep_slave_threads    = 16

binlog_format                  = ROW
default_storage_engine         = InnoDB
innodb_autoinc_lock_mode       = 2
innodb_locks_unsafe_for_binlog = 1
innodb_flush_log_at_trx_commit = 2

# * IMPORTANT: Additional settings that can override those from this file!
#   The files must end with '.cnf', otherwise they'll be ignored.
!includedir /etc/mysql/conf.d/

I also realize that the xtrabackup needs access rights made available so I have also made a GRANT entry for that user from the remote servers and tested via manual login, still no dice.

Any ideas?

I have a command line process that is ran by my ruby on rails application to use the ImageMagick 'convert' command to convert a PDF into multiple PNGs. The problem is that when I run the command it takes all the memory on my VPS (512MB) and renders many other functions near useless (web server, ssh etc).

My convert command is this:

convert -density 288 ./document.pdf -resize 25% ./pages/page_%03d.png

I've tried to use ulimit to limit the memory this process can consume to roughly 15% of memory which on my 512MB VPS is roughly 76,800 kbytes. If I run the above command with ulimit as seen below the process initially spikes with no limit taking up to 80% of memory and renders other functions of the server useless due to lag. After a while the process comes down to under 10% (usually 4-8% of memory) but all the other server's functionality is still slow.

Note my numbers about how much memory the process is consuming are coming from top.

bash -c 'ulimit -m 76800; convert -density 288 ./document.pdf -resize 25% ./pages/page_%03d.png'

Does anyone have any ideas on how I can execute this command without it consuming all my memory?

Thanks!

I have a Ruby On Rails application running on Nginx which is serving out MP3s using JW player. I need to be able to set the start time and duration for playlist items. From what I can tell to do this I need be streaming the MP3 files. How can I setup Nginx to do this?

I'm running an RRD dump from Citrix XenServer and I'm parsing the XML for import into another system. I know that within the element there is a element which is a timestamp, but when I parse the integer into a Time sometimes it is in the future. Is RRD interpolating what the values will be?

Here's an example of the XML: http://pastie.org/1541903

How can I setup an nginx proxy_pass directive that will also include HTTP Basic authentication information sent to the proxy host?

This is an example of the URL I need to proxy to:

http://username:[email protected]/export?uuid=1234567890

The end goal is to allow 1 server present files from another server (the one we're proxying to) without exposing the URI of the proxy server. I have this working 90% correct now from following the Nginx config found here:

http://kovyrin.net/2010/07/24/nginx-fu-x-accel-redirect-remote/

I just need to add in the HTTP Basic authentication to send to the proxy server

I have a pool of XenServer hosts running the Free version of XenServer 5.6 FP1. I was wondering if I change the network backend to use Open vSwitch if I can specify ACLs on individual network VIFs without needing to use the DVS appliance (distributed virtual switch) which requires an Advanced License or higher.

Basically I'm looking for a way to isolate VMs on my network so that if a user had root access on the command line they couldn't access other servers they should not be able to (without using a VLAN).

I'm trying to use xvpviewer (based on TightVNC) to VNC into my VMs running on Citirx XenServer. There are a couple of caveats required with trusting the certificate from XenServer which I've got working. Essentially I'm trying to convert the java command below (which works on the command line to launch VncViewer) for use in an applet that can be accessed via HTML page.

java -Djavax.net.ssl.trustStore=/tmp/kimo.jks -Xmx64m -jar VncViewer.jar HOST "/console?ref=OpaqueRef:141f4204-2240-4627-69c6-a0c7d9898e6a&session_id=OpaqueRef:91a483c4-bc40-3bb0-121c-93f2f89acc3c" PORT 443 PROXYHOST1 192.168.0.5 PROXYPORT1 443 SocketFactory "HTTPSConnectSocketFactory"

I know I can put the HOST, PORT etc arguments into param tags for the applet but I'm not sure how to apply the two initial argments.

I have a network of VMs each one having a LAN IP address and a public IP address. They each have a 1:1 NAT map for public access via the public IP for HTTP, SSH, etc. I'm trying to figure out a way to restrict the LAN IPs from communicate to each other, but there are some cases where a group of LAN IPs will need to communicate.

I'm using pfSense as a firewall / router on a 192.168.0.0/24 configuration.
It seems like I could assign each VM it's own subnet and add a static route to the firewall for that VM to get back to the firewall for internet access and other firewall rules. Is that right?

I assigned 1 VM with:

address 192.168.1.2
netmask 255.255.255.254
gateway 192.168.1.1

Then added a static route on the FW's LAN interface using 192.168.1.0/30 as the destination network and 192.168.1.1 as the gateway.

Nothing appears to be working, so anyone has any ideas? Please be aware I'm not that familiar with subnets.

Thanks!

I'm looking for a way to be able to track the bandwidth of multiple vhosts on an nginx web server. I'm guessing there is a way that I can set up the log files to output this information and then I can write a script to parse through the log files and add up the file sizes.

If that is the case does anyone know the correct log format, and if there is already a script out there that does this?

I'm setting up a server which other people will have sudo access to via ssh. They can install stuff and make changes as they see fit, however I still need to manage the server, install patches and software updates etc.

I'm wondering if I keep the root password so that I can always do what I need to, can I somehow disable people with sudo from being able to change it with sudo.

Does anyone know if there is a full implementation of the Xen API in ruby floating around out there somewhere? I found this: http://github.com/rubiojr/pangea/tree/master but it is read only and is not fully developed.