So I guess this can go in either Virtual machine group but since I use XEN I posted it here.  I can confirm that using the CrashPlan linux installer it will backup files located on samba network shares. 

 

I have a CentOS6 VM with my photos and other unraid shares mounted in /mnt and the crashplan program backs them up with no problem.  An option for some of you folks that want to backup to a cloud service without trying to install the crashplan package that's floating around.. 

Crashplan is also available to be installed as a Docker container.  I would recommend that approach over a VM for this particular app.

Not a fan of the docker world.  In my eyes its slow and not as stable as I feel it should be.  As a system admin by trade I have more faith in a skinny VM that I have created to handle my applications.  Plus I have a pure hatred of ubuntu which is what seems to be the run away base that people are developing for.  Just a personal preference. 

Not a fan of the docker world.  In my eyes its slow and not as stable as I feel it should be.  As a system admin by trade I have more faith in a skinny VM that I have created to handle my applications.  Plus I have a pure hatred of ubuntu which is what seems to be the run away base that people are developing for.  Just a personal preference.

Docker outperforms KVM and Xen on all metrics and VMware as well.  It consumes less resources than a VM and is easier to setup.  Totally understand hesitancy to supporting something new and different, but your statements regarding slowness aren't really accurate. It may have been how you set it up.  Curious how you came to these conclusions...

 

Oh and Ubuntu isn't the only distro you can use with docker.  There are plenty...

Not a fan of the docker world.  In my eyes its slow and not as stable as I feel it should be.  As a system admin by trade I have more faith in a skinny VM that I have created to handle my applications.  Plus I have a pure hatred of ubuntu which is what seems to be the run away base that people are developing for.  Just a personal preference.

Docker outperforms KVM and Xen on all metrics and VMware as well.  It consumes less resources than a VM and is easier to setup.  Totally understand hesitancy to supporting something new and different, but your statements regarding slowness aren't really accurate. It may have been how you set it up.  Curious how you came to these conclusions...

 

Oh and Ubuntu isn't the only distro you can use with docker.  There are plenty...

 

Sorry I should have quantified what I ment by "slow".  When I experimented with Docker on beta6 the disk io was very slow with the base image I was using on btrfs.  File transfers were also a lot slower then I assumed it would be. 

 

As for resources, I don't think that anyone running unraid with either docker or vm's are too worried about a lack of resources.  Looking at some of these builds going on now with 8 core CPUS running at 3ghz or higher and 16 to 32 gig of ram resources are the last thing we are thinking about.   

 

While Ubuntu isn't the only distro it is the most popular one by far for some reason to base Docker images on.  I don't have a problem with Debian but what Ubuntu has evolved to over to years has pushed me away from it. 

Not a fan of the docker world.  In my eyes its slow and not as stable as I feel it should be.  As a system admin by trade I have more faith in a skinny VM that I have created to handle my applications.  Plus I have a pure hatred of ubuntu which is what seems to be the run away base that people are developing for.  Just a personal preference.

Docker outperforms KVM and Xen on all metrics and VMware as well.  It consumes less resources than a VM and is easier to setup.  Totally understand hesitancy to supporting something new and different, but your statements regarding slowness aren't really accurate. It may have been how you set it up.  Curious how you came to these conclusions...

 

Oh and Ubuntu isn't the only distro you can use with docker.  There are plenty...

 

Sorry I should have quantified what I ment by "slow".  When I experimented with Docker on beta6 the disk io was very slow with the base image I was using on btrfs.  File transfers were also a lot slower then I assumed it would be. 

 

As for resources, I don't think that anyone running unraid with either docker or vm's are too worried about a lack of resources.  Looking at some of these builds going on now with 8 core CPUS running at 3ghz or higher and 16 to 32 gig of ram resources are the last thing we are thinking about.   

 

While Ubuntu isn't the only distro it is the most popular one by far for some reason to base Docker images on.  I don't have a problem with Debian but what Ubuntu has evolved to over to years has pushed me away from it.

 

Hmm, interested if this was a spinning disk or not, what containers you were running, and how you structured your tests.  Clearly there are a majority of our users now using Docker and not reporting these problems.

 

As for memory, check our polls and you'll see that the majority of unRAID users are NOT using as much RAM as you indicate.  Most use 4gb or less.  In addition, many opt for low power consumption processors making it even more critical to be efficient in usage.

 

Lastly, distros shouldn't matter with docker very much for the majority of people.  Its the apps that should matter. You don't have to manage the images really, just the apps.

 

All of that being said, I should also note that definitely are a more advanced user than most and if this is just how you prefer to do things for your own personal sanity, I totally understand that.  I just want to be clear that Docker isn't slower at all by any true tests that have been performed (especially check the report from IBM on Docker vs KVM:  KVM got smoked on all performance metrics)

Not a fan of the docker world.  In my eyes its slow and not as stable as I feel it should be.  As a system admin by trade I have more faith in a skinny VM that I have created to handle my applications.  Plus I have a pure hatred of ubuntu which is what seems to be the run away base that people are developing for.  Just a personal preference.

Docker outperforms KVM and Xen on all metrics and VMware as well.  It consumes less resources than a VM and is easier to setup.  Totally understand hesitancy to supporting something new and different, but your statements regarding slowness aren't really accurate. It may have been how you set it up.  Curious how you came to these conclusions...

 

Oh and Ubuntu isn't the only distro you can use with docker.  There are plenty...

 

Sorry I should have quantified what I ment by "slow".  When I experimented with Docker on beta6 the disk io was very slow with the base image I was using on btrfs.  File transfers were also a lot slower then I assumed it would be. 

 

As for resources, I don't think that anyone running unraid with either docker or vm's are too worried about a lack of resources.  Looking at some of these builds going on now with 8 core CPUS running at 3ghz or higher and 16 to 32 gig of ram resources are the last thing we are thinking about.   

 

While Ubuntu isn't the only distro it is the most popular one by far for some reason to base Docker images on.  I don't have a problem with Debian but what Ubuntu has evolved to over to years has pushed me away from it.

 

Hmm, interested if this was a spinning disk or not, what containers you were running, and how you structured your tests.  Clearly there are a majority of our users now using Docker and not reporting these problems.

 

As for memory, check our polls and you'll see that the majority of unRAID users are NOT using as much RAM as you indicate.  Most use 4gb or less.  In addition, many opt for low power consumption processors making it even more critical to be efficient in usage.

 

Lastly, distros shouldn't matter with docker very much for the majority of people.  Its the apps that should matter. You don't have to manage the images really, just the apps.

 

All of that being said, I should also note that definitely are a more advanced user than most and if this is just how you prefer to do things for your own personal sanity, I totally understand that.  I just want to be clear that Docker isn't slower at all by any true tests that have been performed (especially check the report from IBM on Docker vs KVM:  KVM got smoked on all performance metrics)

 

It was a spinning disk as opposed to an SSD.  I am positive that an SSD would have been a lot faster. 

 

As for the resources I was speaking about, looking through the build thread is where I got my frame of reference.  Just going by whats on the main page I only saw 4 or so builds that were going low power. 

 

I guess what it really boils down to for me at least its about having control over the software that I am trusting all my data too.  Well insofar as the applications that serve it or allow me interact with it.  I am also one of those guys that don't trust the cloud either.  While I was glad that I could get CrashPlan installed to backup my family photos it is my 3rd backup option / location providing for my offsite backup.   

 

 

Not a fan of the docker world.  In my eyes its slow and not as stable as I feel it should be.  As a system admin by trade I have more faith in a skinny VM that I have created to handle my applications.  Plus I have a pure hatred of ubuntu which is what seems to be the run away base that people are developing for.  Just a personal preference.

Docker outperforms KVM and Xen on all metrics and VMware as well.  It consumes less resources than a VM and is easier to setup.  Totally understand hesitancy to supporting something new and different, but your statements regarding slowness aren't really accurate. It may have been how you set it up.  Curious how you came to these conclusions...

 

Oh and Ubuntu isn't the only distro you can use with docker.  There are plenty...

 

Sorry I should have quantified what I ment by "slow".  When I experimented with Docker on beta6 the disk io was very slow with the base image I was using on btrfs.  File transfers were also a lot slower then I assumed it would be. 

 

As for resources, I don't think that anyone running unraid with either docker or vm's are too worried about a lack of resources.  Looking at some of these builds going on now with 8 core CPUS running at 3ghz or higher and 16 to 32 gig of ram resources are the last thing we are thinking about.   

 

While Ubuntu isn't the only distro it is the most popular one by far for some reason to base Docker images on.  I don't have a problem with Debian but what Ubuntu has evolved to over to years has pushed me away from it.

 

Hmm, interested if this was a spinning disk or not, what containers you were running, and how you structured your tests.  Clearly there are a majority of our users now using Docker and not reporting these problems.

 

As for memory, check our polls and you'll see that the majority of unRAID users are NOT using as much RAM as you indicate.  Most use 4gb or less.  In addition, many opt for low power consumption processors making it even more critical to be efficient in usage.

 

Lastly, distros shouldn't matter with docker very much for the majority of people.  Its the apps that should matter. You don't have to manage the images really, just the apps.

 

All of that being said, I should also note that definitely are a more advanced user than most and if this is just how you prefer to do things for your own personal sanity, I totally understand that.  I just want to be clear that Docker isn't slower at all by any true tests that have been performed (especially check the report from IBM on Docker vs KVM:  KVM got smoked on all performance metrics)

 

It was a spinning disk as opposed to an SSD.  I am positive that an SSD would have been a lot faster. 

 

As for the resources I was speaking about, looking through the build thread is where I got my frame of reference.  Just going by whats on the main page I only saw 4 or so builds that were going low power. 

 

I guess what it really boils down to for me at least its about having control over the software that I am trusting all my data too.  Well insofar as the applications that serve it or allow me interact with it.  I am also one of those guys that don't trust the cloud either.  While I was glad that I could get CrashPlan installed to backup my family photos it is my 3rd backup option / location providing for my offsite backup. 

 

Well an SSD would always be faster for sure but now I'm also wondering if your tests were skewed. What device did you format with btrfs in b6?  If it was an array device and that is where you installed docker, and you are comparing Docker performance running on a  spinning disk in the array compared to a VM running on a spinning disk assigned as cache...there's the reason.  Moving a VM to the array device and you'll probably see the same performance issues.

 

As for the build threads, the folks that post those are typically the minority, proudly showing off their new toys ;-). As much as we think that is awesome, there is an even greater silent majority of our customers running on boxes they might not feel worthy of posting here for critique by the big box users ;-)

 

As for trusting the cloud, I'm with yah.  That's why I like docker, even though they host the images, the build for each image can be screened by the user if they really wanted to. Then again, most don't.  But this is much better than a hosted VM which would have to be downloaded and then examined in much more detail before trusting. Then again, if you build your own VM, you inherently trust it.

 

I'm VERY cautious of what I store in the cloud nowadays and I am slowly eliminating my dependence on traditional cloud storage solutions completely (e.g. Dropbox, OneDrive, etc.).  The one nice thing with Crashplan is you don't have to backup to their cloud.  You and a buddy can target each other for backing up.

 

For what it is worth, I've actually visited and met with the guys at Crashplan (Code42 in Minneapolis, MN).  Definitely a good team and they have a great vision.

 

 

 

Not a fan of the docker world.  In my eyes its slow and not as stable as I feel it should be.  As a system admin by trade I have more faith in a skinny VM that I have created to handle my applications.  Plus I have a pure hatred of ubuntu which is what seems to be the run away base that people are developing for.  Just a personal preference.

Docker outperforms KVM and Xen on all metrics and VMware as well.  It consumes less resources than a VM and is easier to setup.  Totally understand hesitancy to supporting something new and different, but your statements regarding slowness aren't really accurate. It may have been how you set it up.  Curious how you came to these conclusions...

 

Oh and Ubuntu isn't the only distro you can use with docker.  There are plenty...

 

Sorry I should have quantified what I ment by "slow".  When I experimented with Docker on beta6 the disk io was very slow with the base image I was using on btrfs.  File transfers were also a lot slower then I assumed it would be. 

 

As for resources, I don't think that anyone running unraid with either docker or vm's are too worried about a lack of resources.  Looking at some of these builds going on now with 8 core CPUS running at 3ghz or higher and 16 to 32 gig of ram resources are the last thing we are thinking about.   

 

While Ubuntu isn't the only distro it is the most popular one by far for some reason to base Docker images on.  I don't have a problem with Debian but what Ubuntu has evolved to over to years has pushed me away from it.

 

Hmm, interested if this was a spinning disk or not, what containers you were running, and how you structured your tests.  Clearly there are a majority of our users now using Docker and not reporting these problems.

 

As for memory, check our polls and you'll see that the majority of unRAID users are NOT using as much RAM as you indicate.  Most use 4gb or less.  In addition, many opt for low power consumption processors making it even more critical to be efficient in usage.

 

Lastly, distros shouldn't matter with docker very much for the majority of people.  Its the apps that should matter. You don't have to manage the images really, just the apps.

 

All of that being said, I should also note that definitely are a more advanced user than most and if this is just how you prefer to do things for your own personal sanity, I totally understand that.  I just want to be clear that Docker isn't slower at all by any true tests that have been performed (especially check the report from IBM on Docker vs KVM:  KVM got smoked on all performance metrics)

 

It was a spinning disk as opposed to an SSD.  I am positive that an SSD would have been a lot faster. 

 

As for the resources I was speaking about, looking through the build thread is where I got my frame of reference.  Just going by whats on the main page I only saw 4 or so builds that were going low power. 

 

I guess what it really boils down to for me at least its about having control over the software that I am trusting all my data too.  Well insofar as the applications that serve it or allow me interact with it.  I am also one of those guys that don't trust the cloud either.  While I was glad that I could get CrashPlan installed to backup my family photos it is my 3rd backup option / location providing for my offsite backup. 

 

Well an SSD would always be faster for sure but now I'm also wondering if your tests were skewed. What device did you format with btrfs in b6?  If it was an array device and that is where you installed docker, and you are comparing Docker performance running on a  spinning disk in the array compared to a VM running on a spinning disk assigned as cache...there's the reason.  Moving a VM to the array device and you'll probably see the same performance issues.

 

As for the build threads, the folks that post those are typically the minority, proudly showing off their new toys ;-). As much as we think that is awesome, there is an even greater silent majority of our customers running on boxes they might not feel worthy of posting here for critique by the big box users ;-)

 

As for trusting the cloud, I'm with yah.  That's why I like docker, even though they host the images, the build for each image can be screened by the user if they really wanted to. Then again, most don't.  But this is much better than a hosted VM which would have to be downloaded and then examined in much more detail before trusting. Then again, if you build your own VM, you inherently trust it.

 

I'm VERY cautious of what I store in the cloud nowadays and I am slowly eliminating my dependence on traditional cloud storage solutions completely (e.g. Dropbox, OneDrive, etc.).  The one nice thing with Crashplan is you don't have to backup to their cloud.  You and a buddy can target each other for backing up.

 

For what it is worth, I've actually visited and met with the guys at Crashplan (Code42 in Minneapolis, MN).  Definitely a good team and they have a great vision.

 

 

I had a simple 500gig sata3 drive outside of the array formatted btrfs.  I didn't use my cache drive.  I pulled down what seemed to be the official unraid docker image to play around with.   

 

There was no scientific method involved just a seat of the pants experience.  Granted my seat might be more or less sensitive then others.  It could have been a btrfs issue  or some other disk issue as well because VM's on btrfs ran horribly slow for me as well.  In my centos vm's I had  iowait times over 10 and system 5 and 10 min loads over 9 with an idle vm.  But I am currently using the same disk formatted ext4 and it seems to perform as expected.

 

 

Not a fan of the docker world.  In my eyes its slow and not as stable as I feel it should be.  As a system admin by trade I have more faith in a skinny VM that I have created to handle my applications.  Plus I have a pure hatred of ubuntu which is what seems to be the run away base that people are developing for.  Just a personal preference.

Docker outperforms KVM and Xen on all metrics and VMware as well.  It consumes less resources than a VM and is easier to setup.  Totally understand hesitancy to supporting something new and different, but your statements regarding slowness aren't really accurate. It may have been how you set it up.  Curious how you came to these conclusions...

 

Oh and Ubuntu isn't the only distro you can use with docker.  There are plenty...

 

Sorry I should have quantified what I ment by "slow".  When I experimented with Docker on beta6 the disk io was very slow with the base image I was using on btrfs.  File transfers were also a lot slower then I assumed it would be. 

 

As for resources, I don't think that anyone running unraid with either docker or vm's are too worried about a lack of resources.  Looking at some of these builds going on now with 8 core CPUS running at 3ghz or higher and 16 to 32 gig of ram resources are the last thing we are thinking about.   

 

While Ubuntu isn't the only distro it is the most popular one by far for some reason to base Docker images on.  I don't have a problem with Debian but what Ubuntu has evolved to over to years has pushed me away from it.

 

Hmm, interested if this was a spinning disk or not, what containers you were running, and how you structured your tests.  Clearly there are a majority of our users now using Docker and not reporting these problems.

 

As for memory, check our polls and you'll see that the majority of unRAID users are NOT using as much RAM as you indicate.  Most use 4gb or less.  In addition, many opt for low power consumption processors making it even more critical to be efficient in usage.

 

Lastly, distros shouldn't matter with docker very much for the majority of people.  Its the apps that should matter. You don't have to manage the images really, just the apps.

 

All of that being said, I should also note that definitely are a more advanced user than most and if this is just how you prefer to do things for your own personal sanity, I totally understand that.  I just want to be clear that Docker isn't slower at all by any true tests that have been performed (especially check the report from IBM on Docker vs KVM:  KVM got smoked on all performance metrics)

 

It was a spinning disk as opposed to an SSD.  I am positive that an SSD would have been a lot faster. 

 

As for the resources I was speaking about, looking through the build thread is where I got my frame of reference.  Just going by whats on the main page I only saw 4 or so builds that were going low power. 

 

I guess what it really boils down to for me at least its about having control over the software that I am trusting all my data too.  Well insofar as the applications that serve it or allow me interact with it.  I am also one of those guys that don't trust the cloud either.  While I was glad that I could get CrashPlan installed to backup my family photos it is my 3rd backup option / location providing for my offsite backup. 

 

Well an SSD would always be faster for sure but now I'm also wondering if your tests were skewed. What device did you format with btrfs in b6?  If it was an array device and that is where you installed docker, and you are comparing Docker performance running on a  spinning disk in the array compared to a VM running on a spinning disk assigned as cache...there's the reason.  Moving a VM to the array device and you'll probably see the same performance issues.

 

As for the build threads, the folks that post those are typically the minority, proudly showing off their new toys ;-). As much as we think that is awesome, there is an even greater silent majority of our customers running on boxes they might not feel worthy of posting here for critique by the big box users ;-)

 

As for trusting the cloud, I'm with yah.  That's why I like docker, even though they host the images, the build for each image can be screened by the user if they really wanted to. Then again, most don't.  But this is much better than a hosted VM which would have to be downloaded and then examined in much more detail before trusting. Then again, if you build your own VM, you inherently trust it.

 

I'm VERY cautious of what I store in the cloud nowadays and I am slowly eliminating my dependence on traditional cloud storage solutions completely (e.g. Dropbox, OneDrive, etc.).  The one nice thing with Crashplan is you don't have to backup to their cloud.  You and a buddy can target each other for backing up.

 

For what it is worth, I've actually visited and met with the guys at Crashplan (Code42 in Minneapolis, MN).  Definitely a good team and they have a great vision.

 

 

I had a simple 500gig sata3 drive outside of the array formatted btrfs.  I didn't use my cache drive.  I pulled down what seemed to be the official unraid docker image to play around with.   

 

There was no scientific method involved just a seat of the pants experience.  Granted my seat might be more or less sensitive then others.  It could have been a btrfs issue  or some other disk issue as well because VM's on btrfs ran horribly slow for me as well.  In my centos vm's I had  iowait times over 10 and system 5 and 10 min loads over 9 with an idle vm.  But I am currently using the same disk formatted ext4 and it seems to perform as expected.

Weird.  Well one day I hope to win you over enough to give Docker another shot.  As a fellow technologist, I think you'd get a real kick out of how truly and ridiculously efficient Docker is compared to traditional VMs.  Again, today understand why you're doing what you're doing now, and not trying to hijack this thread to "convert" you, but wanted to better understand how you came to your initial conclusions about Docker.

 

I will readily admit I'm a docker fanboy, but I also know that Docker is a lot younger technology than traditional virtualization so when I see feedback like yours, I want to dig deeper...

I am actually in the process of preclearing drive.  Will docker run on XFS or will it only run on the btrfs? 

I am actually in the process of preclearing drive.  Will docker run on XFS or will it only run on the btrfs?

Any.  We changed how this works since beta 6.  Now it installs to a vdisk and you specify the location on the webGUI.  It'll ask you for a size to give it.  You can always grow the # but you can't shrink it.

Crashplan is also available to be installed as a Docker container.  I would recommend that approach over a VM for this particular app.

 

Could you elaborate on why you would run Crashplan as a Docker container?  I'm currently running it in a Ubuntu VM under XEN (14.04 LTS Server) backing up to a user share. 

Crashplan is also available to be installed as a Docker container.  I would recommend that approach over a VM for this particular app.

 

Could you elaborate on why you would run Crashplan as a Docker container?  I'm currently running it in a Ubuntu VM under XEN (14.04 LTS Server) backing up to a user share.

Simpler to setup initially, doesn't require a full VM with hardware emulation to operate, and consumes less space in both disk and memory.

 

I would add that if you just need crashplan then maybe a docker is the way to go.  I haven't gotten on the docker wagon yet altho I have tried it out.  Personally I run a CentOS VM with crashplan. But I run other services off Centos as well so it works for me.  I also run a debian VM that houses some stuff that I could fight to run on Cent but I too the easy road on that..