• 0 Posts
  • 31 Comments
Joined 1 year ago
cake
Cake day: June 23rd, 2023

help-circle
  • This is the method I use in your scenario, OP. You can use Folder2iso to get the files in that you need. If the OS has official VMware tools, you can also mount the VMware Tools ISO straight from workstation into the VM and this will give you the clipboard service so you can copy and paste files between the host and VM, if this scenario is permitted within your isolation needs.

    Otherwise, go the ISO route. You just can’t bring stuff out of the VM back to the host is all.



  • Its very much still needed and heavily utilised in the enterprise world. Volume size is usually the lowest priority when it comes to arrays, redundancy and IOPS (the amount of concurrent transactions to the storage) is typically the priority. The exception here would be backup and archive storage, where IOPS is less important and volume size is more important.

    As far as replacing sectors goes, I’ve never heard of this and I might just be ignorant on the subject but as far as I know you can’t “replace” a bad sector. Only mark it as bad and not use it, and whatever was there before is gone. This has existed since HDD days. This is also why we use RAID - parity across disks to protect data.

    Generally production storage will be in RAID-10, and backup/archive storage in RAID-6 or in some cases RAID-60 but I’m personally not a fan.

    You also would consider how many disks are in the volume because there is a sweet spot. Too many disks = higher likelihood of total array failure due to simultaneous disk failures and more data loss in the event it does, but too few disks and you won’t have good redundancy, capacity or performance either (depending on RAID level).

    The biggest change I see in RAID these days is moving away from hardware RAID cards and into software-based solutions like Microsoft Storage Spaces, md, ZFS and similar. These all have their own way of doing things and some can even synchronise the data with other hosts.

    Hope this helps!









  • TIL about Fedora, last I knew it was a rolling bleeding edge OS. Clearly lots of movement in the Red Hat camp.

    As for gaming, drivers were not the problem for me. Getting games to run with ease was. On OpenSUSE, I just install Steam, enable Proton and basically go at that point. Red Hat was non-trivial to do this. Could be a skill issue, but I had a better time getting going with OpenSUSE TW.


  • Sort of, OpenSUSE Tumbleweed. I started on OpenSUSE Leap but had issues getting things like GPU and Steam working. Red Hat was also a non-starter because of the lack of gaming functionality.

    TW works great for gaming and the enterprise features I care about (like domain joining) work out of the box. Its certainly harder to set up than something more geared towards home use (typically one of the various the downstreams of Debian or Arch) but that doesn’t bother me.


  • Second to this - for what its worth (and I may be tarred and feathered for saying this here), I prefer commercial software for my backups.

    I’ve used many, including:

    • Acronis
    • Arcserve UDP
    • Datto
    • Storagecraft ShadowProtect
    • Unitrends Enterprise Backup (pre-Kaseya, RIP)
    • Veeam B&R
    • Veritas Backup Exec

    What was important to me was:

    • Global (not inline) deduplication to disk storage
    • Agent-less backup for VMware/Hyper-V
    • Tape support with direct granular restore
    • Ability to have multiple destinations on a backup job (e.g. disk to disk to tape)
    • Encryption
    • Easy to set up
    • Easy to make changes (GUI)
    • Easy to diagnose
    • Not having to faff about with it and have it be the one thing in my lab that just works

    Believe it or not, I landed on Backup Exec. Veeam was the only other one to even get close. I’ve been using BE for years now and it has never skipped a beat.

    This most likely isn’t the solution for you, but I’m mentioning it just so you can get a feel for the sort of considerations I made when deciding how my setup would work.


  • As others have mentioned its important to highlight the difference between a sync (basically a replica of the source) vs a true backup which is historical data.

    As far as tools goes, if the device is running OMV you might want to start by looking at the options within OMV itself to achieve this. A quick google hinted at a backup plugin that some people seem to be using.

    If you’re going to be replicating to a remote NAS over the Internet, try to use a site-to-site VPN for this and do not expose file sharing services to the internet (for example by port forwarding). Its not safe to do so these days.

    The questions you need to ask first are:

    1. What exactly needs to be backed up? Some of it? All of it?
    2. How much space does the data I need backed up consume? Do I have enough to fit this plus some headroom for retention?
    3. How many backups do I want to retain? And for how long? (For example you might keep 2 weeks of daily backups, 3 months of weekly backups, 1 year of monthly backups)
    4. How feasible is it to run a test restore? How often am I going to do so? (I can’t emphasise test restores enough - your backups are useless if they aren’t restorable)
    5. Do you need/want to encrypt the data at rest?
    6. Does the internet bandwidth between the two locations allow for you to send all the data for a full backup in a reasonable amount of time or are you best to manually seed the data across somehow?

    Once you know that you will be able to determine:

    1. What tool suits your needs
    2. How you will configure the tool
    3. How to set up the interconnects between sites
    4. How to set up the destination NAS

    I hope I haven’t overwhelmed, discouraged or confused you more and feel free to ask as many questions as you need. Protecting your data isn’t fun but it is important and its a good choice you’re making to look into it


  • Back in the day when the self-hosted $10 license existed I was using JIRA Service Desk to do this. As far as ticketing systems go it was very easy to work with and didn’t slow me down too much.

    I know you don’t want a ticket system but I’m just curious what other people will suggest because I’m in the same boat as you.

    Currently I haphazardly use Joplin to take very loose notes and sync them to Nextcloud.

    If you want a very simple option with minimal setup and overhead you could use Joplin to create separate notes for each “part” of your lab and just add a new line with a date, time and summary of the change.

    I do also use SnipeIT to track all my hardware and parts, which allows you to add notes and service history against the hardware asset.

    Other than that, I’m keen to see what everyone else says


  • Servers are a different story but for Desktop, OpenSUSE.

    Because:

    • It’s stable even on their rolling OS (Tumbleweed)
    • Gaming works exceptionally well
    • CUDA works with little effort
    • RPM-based (personal preference)
    • zypper is an excellent package manager and my experience has been better than that of yum/dnf
    • Extensive native packages and 3rd party repos
    • No covert advertising in the OS
    • Minimal (no?) Telemetry
    • Easy to bind to active directory
    • it feels polished and well built
    • I do not have to mess with it to make it work

    Part of my transition from Windows to Linux was that basic tasks like installing software or even the OS itself shouldn’t be a high effort endeavour. I should be able to point to a package file or run a package manager and be able to go about my day without running “make” and working my way through dependency hell.

    I say this as a Linux user of all different flavours for well over 15 years who has a deep love for what it brings to the table. If we want it to be common place with non-IT folks, it needs to work and it needs to be simple to use.


  • Power

    • 2x feeds into the rack (same circuit but we’ll work on that)
    • Eaton 2000VA double conversion UPS on Feed A
    • APC 1500VA line interactive UPS on Feed B (bypassed, replacing it with another double conversion 2kVA eventually)

    Network

    • 2x Dell N2048P, stacked (potentially getting replaced with 2x stacked Cisco 9300)
    • FortiGate firewall
    • 1000/50 FTTP primary Internet link
    • 4G backup Internet link using a different Telco (the dream is to replace this with Starlink)

    Storage

    • Synology 4-bay NAS with 4x4TB in RAID-10 (for overflow storage from Virtual SAN cluster)
    • HP MSL2024 8GB Fiber Channel LTO5 Tape autoloader for off-site backup

    Compute

    • Dell R520 running VMware ESX for Production (2x Xeon E5-2450L, 80GB DDR3, 4x500GB SSD RAID-10 for Virtual SAN, 1x10TB SATA “scratch” disk, 2x10G fibre storage NICs, 2x1G copper NICs for VM traffic)
    • Dell R330 running VMware ESX for backups and DR (1x Xeon E3-1270v5, 32GB DDR4, 2x512GB SSD RAID-1, 2x4TB HDD RAID-1, 8G FC card for tape library)

    A second prod host will join the R520 soon to add some redundancy and mirror the Virtual SAN.

    All VMs are backed up and kept in an encrypted on-site data store for at least 4 weeks. They’re duplicated to tape (encrypted) once a month and taken off site. Those are kept for 1 year minimum. Cloud backup storage will never replace tape in my setup.

    Services

    As far as “public facing” goes, the list is very short:

    Though I do run around 30-40 services all up on this setup (not including actual non-prod lab things that are on other servers or various SBCs around the place).

    If I had unlimited free electricity and no functioning ears I’d be using my Cisco UCS chassis and Nexus 5K switch/fabric extenders. But it just isn’t meant to be (for now, haha).



  • Worked for an MSP, we had a large storage array which was our cloud backup repository for all of our clients. It locked up and was doing this semi-regularly, so we decided to run an “OS reinstall”. Basically these things install the OS across all of the disks, on a separate partition to where the data lives. “OS Reinstall” clones the OS from the flash drive plugged into the mainboard back to all the disks and retains all configuration and data. “Factory default”, however, does not.

    This array was particularly… special… In that you booted it up, held a paperclip into the reset pin, and the LEDs would flash a pattern to let you know you’re in the boot menu. You click the pin to move through the boot menu options, each time you click it the lights flash a different pattern to tell you which option is selected. First option was normal boot, second or third was OS reinstall, the very next option was factory default.

    I head into the data centre. I had the manual, I watched those lights like a hawk and verified the “OS reinstall” LED flash pattern matched up, then I held the pin in for a few seconds to select the option.

    All the disks lit up, away we go. 10 minutes pass. Nothing. Not responding on its interface. 15 minutes. 20 minutes, I start sweating. I plug directly into the NIC and head to the default IP filled with dread. It loads. I enter the default password, it works.

    There staring back at me: “0B of 45TB used”.

    Fuck.

    This was in the days where 50M fibre was rare and most clients had 1-20M ADSL. Yes, asymmetric. We had to send guys out as far as 3 hour trips with portable hard disks to re-seed the backups over a painful 30ish days of re-ingesting them into the NAS.

    The worst part? Years later I discovered that, completely undocumented, you can plug a VGA cable in and you get a text menu on the screen that shows you which option you have selected.

    I (somehow) did not get fired.


  • Because prospective customers get shy when the browser says that your site is “insecure” Because it factually is insecure. It is not encrypted and trivial to inspect.

    Because it makes for better google ranking. No, in this day and age it is permission to play. Firefox has a built in feature to only load HTTPS sites, which I have enabled. This has nothing to do with Google. Your issue is with expensive CAs, to which there is a free solution (Let’s Encrypt). Not HTTPS itself.

    So there you go. Mob hype and googlian dictatorship. Incorrect. It is a matter of safety and security and a trivial thing to implement. You are free to not use HTTPS if you want, just as people are free to not consume your service if you don’t.

    Calling it a “dictatorship” is hyperbole and demonstrates that you clearly have no idea what you’re talking about and won’t listen to people that do.