Sysadmin Mega Thread

This shit is so annoying. So much security through obsfucation in routers.

Working on a Ubiquiti EdgeOS guide.

1 Like

So, i made a script. For vCenter VM startup, in a better way than dumb low-medium-high priorities. Has a per-vm delay you can configure on a tag by tag basis - so you can avoid hammering the fuck out of your disk subsystem by starting everything at once. Also, allows you to do stuff like ensure VMs start after other VMs they depend on :smiley:

Dirty. No logs. No exception handling (i’m going to add that, but this actually works as is). But will start VMs based on tag and a per-vm delay for VMs within a tag group. Oh yeah, it needs vmware powercli (for powershell) to be installed. Works on Mac or Linux though.

$VMTagInfo = import-csv cluster-start-tags.csv
$SortedTagInfo = $VMTagInfo | sort-object -Property Order
connect-viserver your-vcenter

"Tags found for VM startup"
$SortedTagInfo


"iterating through tiers found... "

foreach ($tag in $SortedTagInfo) {

  $VMList = get-vm -tag $Tag.Name
  foreach ($VM in $VMList) {
    "Starting VM: " + $VM.Name
    start-vm $VM.Name
    "Delaying: " + $Tag.PerVMDelay
    start-sleep -s $Tag.PerVMDelay

  }
}

# Clean up
disconnect-viserver

The example CSV file
Name,Order,PerVMDelay
JRTEST,1,10

edit:
just noticed there’s a “wait-tools” cmdlet, which would enable me to wait for VMware tools in a guest VM to start. Could be handy…

1 Like

I’ve found that dedup works best when compression is off, using the default chunk size

Lol thanks!

I had the idea that if you look at it all day that warm colors would help keep your mood positive.

2 Likes

That looks pretty slick. You guys are doing an in-house ticketing system rather than third party or vendor?

I’m just making it for fun.

That said my boss does know about what I’m making.

1 Like

One thing that always bugs me about the ticket systems that are out there is how over-complicated and noisy the interfaces always are.

You know something I really liked in ConnectWise/LabTech that I’ve never seen anywhere else is the SME blocks/groups.

You could create profiles for users based on their strengths and then set up an alert or queue for that SME group.

So, systems engineers, network engineers, application support, etc. Things like that.

Not trying to influence your development efforts (oh I am) but mentioning noisy/irrelevant interfaces reminded me of that.

2 Likes

oh that’s exactly what the “My Queues” section is going to be.

All queues are going to be visible via the ‘All Queues’ button so analysts can see where stuff is assigned at a glance.

I am going to add a section under settings which allows one to select which queues they want to appear under the ‘My Queues’ tab.

3 Likes

Yeah idk why priority queues are so hard for ticketing systems

1 Like

Has anyone ever configured iscsi in bios? Pros/cons/use-case?

I always see it there and I’m like “huh” and don’t touch it.

Nope. use case = boot from SAN

1 Like

Hey folks, I’m back #goldenhandcuffs

Life as a dev really sucks. You have to drive all the time and get massive pay cuts. SysOps 4 Lyfe :wink::wink::wink:

1 Like

Isn’t that backwards? Ops always has to drive to the datacenter to plug in usb sticks and replace fans and whatever, while devs get to work over ssh while they sip mojhitos on the beach…

5 Likes

Mayhap. I could just be fortunate at my current gig. Everything is SDN or VM or AWS services lol.

#bestofbothworlds?

2 Likes

Yeah dude. Best of both worlds.

1 Like

God bless expect scripts

1 Like

Virtual media, out of band management and vendor support agreements = i don’t do hardware :smiley:

Check this out. When someone runs the cp command against multiple directories which are on an NFS mount point to another local directory, the cp command into uninterruptible sleep.

Any idea why this happens?

probably because it’s monday.

4 Likes