Goro's questions

I constantly get the unable to write to * error, even though I cleaned up a bunch of space.

df thinks my disk is 28G in size and 100% full after I deleted a lot of unused files:

$ df -h
Filesystem      Size  Used Avail Use% Mounted on
/dev/sda1        28G   28G     0 100% /

The total size of / is now:

$ sudo du -hs /
484G    /

And here's my physical system info:

$ mount
/dev/sda1 on / type ext4 (rw,errors=remount-ro)

$ sudo parted -l
Model: ATA ST1000DM003-1CH1 (scsi)
Disk /dev/sda: 1000GB
Sector size (logical/physical): 512B/4096B
Partition Table: msdos

Number  Start   End     Size    Type      File system     Flags
 1      1049kB  30.0GB  30.0GB  primary   ext4            boot
 2      30.0GB  100GB   70.0GB  primary   ext4
 3      100GB   999GB   899GB   extended
 5      100GB   999GB   899GB   logical   ext4
 4      999GB   1000GB  1285MB  primary   linux-swap(v1)

Update:

The issue was a runaway log file that filled up /dev/sda1. I suppose that what caused df to not show stats for anything other than /dev/sda1 is the fact that they weren't mounted (as output of mount shows above. As to why that happened I have no idea, but after freeing some space and restart, the system was able to mount everything properly.

I have some random files that I would like to collect and forward to my logging server. These are applications that don't really support GELF, so I am trying to forward these files with rsyslog:

# Apache access log
input(type="imfile" File="/var/log/misc/blah.log"
Tag="Apache Access Log"
StateFile="statefile1")


*.* @@log.ospreyreach.com:12514

Some issues/questions:

1. This forwards all the syslog files. How can I specify only certain specific files to get forwarded?
2. This does not seem to collect any data from the file I defined. I see regular syslog messages popping up in my graylog server, but not that file.

I am running my DB server in amazon cloud, and I have the database files on a separate EBS volume. When it comes to backup/restore operations, I found it infinitely simpler to just do a filesystem-level backup rather than sql dump because I can create backups and restore from them almost instantly.

Are there any possible issues I may be overlooking if I stick to using solely filesystem-level backups?

Running PostgreSQL 9.1 (updating to 9.3 later this year) on Ubuntu 12.04

I have 3x 4TB drives for a ZRAID1 config (single parity drive)

After running

zpool create data zraid1 sdb sdc sdd

I have a volume that's 10.9T in size, which is what I would expect from a stripe configuration, not a RAID5 configuration. zpool status claims it is a raidz1-0 config (the 0 makes sense, this being the only pool on this machine).

Why is my size not 7.2T ?

Figured it out: zpool list shows total of the pool, not the resultant volume size. df -h shows that correctly as 7.2T.

I think we've reached the limit of usability of celery events and celery fllower. Is there a good tool for monitoring status of Celery running under RabbitMQ?

In particular, I would like something to raise an alarm when:

• My queues are backing up
• Consumers register or de-register

It would be also good to know:

• How many tasks executed when - some sort of historical metric logs
• Average task duration, etc

This is for a database (Postgre) VM, whose data is stored on a separate EBS volume. Can I make it so that the same EBS volume is mounted automatically (this is assuming that no other instance is using that EBS).

The reason for this would be that the db would have its data files available when it starts up.

I am looking at creating a PostgreSQL database image (to run on Amazon EC2) and I have heard good things about XFS.

However, since ultimately you are running in a VM, the gain of XFS on EC2 might not be as great as on a barebone system.

Am I going to see a significant improvement with XFS on EC2?

I have a database backup of a MySQL database that is about 250MB. I made it with mysqldump

I was trying to load it on another server like so:

mysql -u xxxx -pxxxx data_mirror < dbdump.sql

I was not satisfied with this process because

1. It did not give me any feedback on the (lengthy) progress
2. It failed with a "Server went away" message

What method would you recommend for backing up and restoring large MySQL database?

I am doing this on Windows 7 - based servers.

I need to monitor traffic coming in and out of my server. Knowing the total in/out bandwidth is good, but ideally I would like to know how much traffic was sent to what address (we are talking about static IP addresses).

Is there a tool that would allow me to track bandwidth usage by destination/source?

Alternatively, can you recommend a good library or API for implementing such a functionality?

Note: At this stage I am looking for a Windows-based tool, but I would like to know of options available for any platform.

Thank you,

We are evaluating hosting our application on Amazon EC2 service. While I am confident that a Small instance (1.7GB RAM) will be sufficient for our web server (at this point), I am not sure about requirements for the MySQL server:

• Would it make sense to put the MySQL server on the same image as our main server? (to save costs)

• I suppose The database will be constantly updated and queried from multiple sources other than the web server, up to 100,000 queries a day / 5000 updates a day. If deploying via RDS, should the database be deployed on an image that has more resources than a "small" instance?

• Would it ever make sense to host a MySQL database on EC2 but not via RDS?

Thanks,

We are evaluating hosting our application on Amazon EC2 service. While I am confident that a Small instance (1.7GB RAM) will be sufficient for our web server (at this point), I am not sure about requirements for the MySQL server:

Would it make sense to put the MySQL server on the same image as our main server? (to save costs)

I suppose The database will be constantly updated and queried from multiple sources other than the web server, up to 100,000 queries a day / 5000 updates a day. If deploying via RDS, should the database be deployed on an image that has more resources than a "small" instance?

Would it ever make sense to host a MySQL database on EC2 but not via RDS?

Thanks,

I am running a Windows 7 installation, and I want to be able to run an application when there is no user logged in. The reason for this is that my application is a monitoring tool, and we want to be able to monitor 24/7 even if the computer reboots. From what I've gathered on the net, there are two ways of doing this:

1. Create a service and run it under an Administrator account

2. Use windows group policies to set up a logon script: This is where things get a bit unclear, as people seem to confuse running a program in the background without a user logged in with running a program just after user has logged in. If I setup my program to run in this mode, will it run if no one logs in?

Can you comment on either #1, #2, which one you would recommend and why? Maybe there is another way I do not know about?

Thank you for the help,

My application is configured to send out email notifications when an event happens. Sometimes this is fairly often (every minute), and every email has an attachment (up to 6-10 attachments), so each email can be pretty hefty, but not ridiculously so (500k per email is the maximum we would ever reach).

The problem we run into is that the emails will get delayed. They will still be delivered, but there will be a delay in there. The strange part about the delay is that it is very consistent (e.g. all emails arrive 1 hour late).

I thought this was a problem with my email server, so I moved my mail service to a (paid) virtual server with lots of memory, but that didnt seem to help (and I have plenty of resources left).

Does anyone have any insights into what might be factors here, or what kind of options do I have? Is moving to a paid SMTP service likely to solve this problem?

Thank you,

I am having a problem with starting a Ruby On Rails application.

When I type

mtr start <application_name>

It will take a long time and fail with the following message:

Error Executing Command.

But the status of the application will show as "running", only that I will immediately get a HTTP 502 error when trying to access it.

Thanks,