Originally Posted by
marcbo
Hi. Thanks for taking the time.
People took the time when I was learning. Just paying back. Please do the same, if you can.
Originally Posted by
marcbo
Maybe that was overkill. I didn't do it. I'm not sure what you were trying to say at the end of that sentence...
I don't remember either. Probably something about using other, available, environment variables or symbolic links. That's a common solution for having data files outside your HOME - setup symbolic links to the other location. This is pretty common for those directories like ~/Music/, ~/Videos/ ~/Pictures/, ~/Documents/ that have been forced onto people by DEs. Or you can stop using a DE completely and just use a window manager. I use fvwm.
Originally Posted by
marcbo
This I did not think about and I did not know. So I changed that backup option to another location on my SSD within my Host OS.
A backup **must** be on different physical storage to count at all. SSDs also wear out, especially if they are nearly full. I've had 2 fail in less than 3 yrs of use. Now I have some new, well regarded, SSDs, but use less than 50% of the claimed storage. That should provide plenty of wear levelling options for the SSD controller. I've read that 20% unused is sufficient to drastically extend the SSD life. OTOH, for backups a $25 spinning USB3 disk is much more cost effective. Better if it has an external power supply so it won't be really slow.
Originally Posted by
marcbo
I'm not sure about this:
If ~ equals /home, then what I meant by ~/marc is /home/marc.(A little confused about this.)
~/marc/ is a directory inside HOME called "marc".
~marc/ is the home directory for a userid "marc" - that any other userid can use to access in that manner.
~/ is the home directory for the current userid, which might be marc or pete or steve.
The ~ ... asks the shell to use the getent passwd $USER command to fill in the correct directory from the password DB. It doesn't assume /home/{something}, which is only used for non-corporate setups. Well, if a ~/ is used, it simply fills in the environment variable for $HOME, but effectively that works in the same way because login pulls the data using getent to set HOME correctly.
Originally Posted by
marcbo
I'm just a loner over here. I don't have a server, except for my little laptop. I do keep multiple backup copies though. 7 in total. So if something goes wrong on day 5, I still have backups 1 through 4 to fall back on. Luckybackup is basically a copy or sync program. I'm not an expert with backups...
Anyways, you are doing pretty well having backups and multiple versions of it. If that works for you, FANTASTIC.
A real versioned backup tool can store 30-60-90 days of backups for very little over the storage required to store just 1 backup. A versioned backup tool will managed the versions for you. The backup tool I like is rdiff-backup. The last backup appears as a mirror, so getting 1 file back is a cp. It is only the older "diffs" where using the tool to get files restored is helpful. The diffs are stored as diff.gz files, so we don't need the backup tool for restoring the backup from 17 days ago, though it is really helpful.
Anyways, I've posted so much here about rdiff-backup that interested people can find those threads, see some issues and some great things about it. Every Sunday morning, I have my backup server make a report about the backups for every system. Here's the report for my main desktop system:
Code:
Time Size Cumulative size
-----------------------------------------------------------------------------
Sun Jun 9 01:15:06 2019 6.26 GB 6.26 GB (current mirror)
Sat Jun 8 01:15:05 2019 68.9 MB 6.33 GB
Fri Jun 7 01:15:06 2019 12.0 MB 6.34 GB
Thu Jun 6 01:15:05 2019 10.9 MB 6.35 GB
Wed Jun 5 01:15:05 2019 15.7 MB 6.37 GB
Tue Jun 4 01:15:05 2019 14.0 MB 6.38 GB
Mon Jun 3 01:15:05 2019 63.0 MB 6.44 GB
...
Tue Apr 16 01:15:08 2019 12.6 MB 7.41 GB
Mon Apr 15 01:15:07 2019 17.1 MB 7.42 GB
Sun Apr 14 01:15:09 2019 24.0 MB 7.45 GB
Sat Apr 13 01:15:09 2019 14.3 MB 7.46 GB
Fri Apr 12 01:15:09 2019 11.3 MB 7.47 GB
Thu Apr 11 01:15:07 2019 11.0 MB 7.48 GB
So, 60 days of versioned backups for 6.26G of files only needs 7.48G of storage. I can assure you that this is sufficient for my to restore the computer in about 30-45 minutes with all the data, all the programs, and all the settings on it at the time of the last backup.
Code:
$ df -h
Filesystem Size Used Avail Use% Mounted on
/dev/vda1 20G 16G 2.6G 86% /
is the total storage on that machine - OS, my files, my settings, etc. I keep large files elsewhere. I don't backup the installed OS, just the metadata about the installed OS so I can put it back quickly. Oh - and my desktop runs inside a KVM virtual machine which is accessible from anywhere (within reason) on the internet through either a VPN (I run a VPN server at home) or ssh connection.
Originally Posted by
marcbo
Of course, I don't know your beefs with that company, but I have similar feelings/suspicions. I don't like working with, and usually don't work with, any company that blocks access to it's instructional Wiki simply because you're using Tor. I personally think that's a red flag.
Around 2010 they gave all their paying clients whiplash by altering the way license costs were determined and their marketing department claimed it would be "better" for almost all their clients. It was going to cost our clients 2-3x more every year. For a few months we had to freak out our clients about the new pricing. We lost a few clients over it who we'd just finished moving to VMware 1 month prior to the new price announcement. I'd been using a few other hypervisors for a while, but we preferred the safety of VMware ESX and the old pricing wasn't THAT bad. Over the next 6 months, we migrated a number of clients to KVM+libvirt and never regretted it. I still use KVM + libvirt here, just a more advanced setup with more flexible storage, faster systems, and more VMs on more physical systems.
Originally Posted by
marcbo
To tell you the truth, I don't know. Opensource vs proprietary is usually at the top of my list in considerations. I suppose after going around in circles with Virtualbox, which drove me crazy, VMware just sort of worked. I'm not an expert at Virtualization either. But I will spend some time today looking into KVM. I'm using an older Thinkpad, which I love, and so I don't have to worry too much about graphic performance as it doesn't have a dedicated GPU that I know of.
QEMU+kvm can used vmdx files, or whatever storage format the VMware tools you have are using. Just need to make the virtual machine virtual hardware closes enough to what VMware is presenting. For Windows guests, it is complicated. For Linux VMs, it is pretty trivial to move between hypervisors. qemu-img can be used to change from one storage file type to another as you like. There is much to like about qcow2, though I only have a few VMs using that. Most of my VMs use LVM LVs presented as block devices to the VM. Probably not something a non-pro would do. Crazy flexible.
Hopefully, this will be helpful too. Take what is good for you and don't worry if nothing is. That's fine too. We're all in different places.
Bookmarks