I can't even
-
@scottalanmiller said in I can't even:
Ever have your boss need you to do a little project for him and it involves just typing two words into a file?
Like, how incompetent can you be?
Best part is...
I'm the boss in this story. lol
@valentina is very long suffering.
@obsolesce said in I can't even:
@scottalanmiller said in I can't even:
Ever have your boss need you to do a little project for him and it involves just typing two words into a file?
Like, how incompetent can you be?
Best part is...
I'm the boss in this story. lol
@valentina is very long suffering.
That's not necessarily incompetence, it's delegation.
@pmoncho said in I can't even:
@scottalanmiller said in I can't even:
Ever have your boss need you to do a little project for him and it involves just typing two words into a file?
Like, how incompetent can you be?
Best part is...
I'm the boss in this story. lol
@valentina is very long suffering.
Sometimes it pays to be the boss!
What is more impressive is when you can delegate up! Getting your boss to do something that you could do is awesome.
-
My manager (IT Manager) messaged me a screenshot of speedtest.net results from one of our DC's that showed there is 33 ms ping times, and that I should look into this. I login the DC, pull up command line, ran a continuous ping to 8.8.8.8 with solid 3ms ping times. I sent him a screenshot of the command prompt and said ping times look solid to me, and speedtest.net does not give accurate results of network latency from a computer as the built-in ping utility from windows.
This is basic troubleshooting, and is just sad.
-
Saw today someone went into their SAN and.... deleted disks from their running RAID array. jajaja.
This is one of the core problems with SANs....
- They deployed a known fragile SAN (SC series) that is insanely more dangerous than a normal server.
- They are using a SAN instead of proper production storage.
- They have "data center" staff managing their really complex and confusing storage instead of storage professionals who know what is going on and how it is used.
Given the three whopping mistakes above, it follows the pattern that human error has led to a degraded and irreparable RAID array on this "infallible" SAN system. It's the problem of "bad decisions beget bad decisions." People installing SANs that don't make sense are also the most likely people to screw up the deployment and managing it wrong and lose data from human error. It's a compounding problem.
-
@Fredtx said in I can't even:
I sent him a screenshot of the command prompt and said ping times look solid to me, and speedtest.net does not give accurate results of network latency from a computer as the built-in ping utility from windows.
Sort of. The ping times ARE accurate. And SpeedTest shows to where they are being sent.
What happened there is that you tested one specifically fast and near server, and he tested and arbitrarily distant server. It's not that his number is more or less accurate than yours, they are both accurate. It is that you are measuring two different things.
What was lacking was a solid definition of what was being sought. A proper network latency test is done to the nearest router, not to any server. Even in your test, the connection is arbitrarily far and can be wrong - assuming the question is "what is our WAN latency". But if the question is "what is our Google DNS latency" then your answer is correct. If the question is "what is the latency to a server selected by SpeedTest" then the manager's test is correct.
-
I've always used the ping command to troubleshoot latency, and view active spikes, and packet loss. The only time I really use Speedtest is when testing bandwidth. Tracert is another command I use for troubleshooting latency as I can easily show the ISP the issue is on their end, which most of the time it is.
-
@Fredtx said in I can't even:
I've always used the ping command to troubleshoot latency, and view active spikes, and packet loss.
For anything ping/traceroute related, the tool you want is
mtrnotping. It provides a much better picture of what you need to know if you use the right--orderoptions.This gives you the standard loss, delay, etc., plus jitter:
mtr 1.1.1.1 --order LSRDNABWVJMXI
But as @scottalanmiller said, you need to be specific about what you are wanting tested.
-
I setup a self-hosted librespeed instance at my prev employer to deal with all the WFH internet foolishness. Then you know that you're testing a known quantity. I'd also managed to script something with their client-side cli and TRMM that would allow the team to remotely speed test without user knowledge or intervention.
-
@JaredBusch said in I can't even:
For anything ping/traceroute related, the tool you want is mtr not ping. It provides a much better picture of what you need to know if you use the right --order options.
Nice. I never knew about that tool. I will have to check it out
-
Today a customer asked us to "return their software and CALs for a refund". The reason given? Because they plan to pirate the software instead so they don't want to have paid for it.
I can't wait to hear what the vendor says to that logic for returning the software.
-
Customer: I don't know enough about computers to buy a server on my own, I need an IT firm to order it for me.
Also Customer: I don't need to pay for IT services for installing my hypervisor, operating systems, backups, application or migrating data - I think that I can do that on my own.
Analogy: I don't know enough to buy a car on my own, but I think that I can work on one without needing a mechanic.
-
@scottalanmiller
Analogy: I crash car into a tree, blame mechanic. -
@scottalanmiller All the time.
-
@scottalanmiller Wow
-
Talking to a customer who tells us that they have a two year old server "so it is time to replace it." Um, what? Who replaces two year old hardware.
Hardware...
Intel 4th Gen Desktop CPU
8GB RAM
Spinning Rust
Windows 10OMG, it's a decade old desktop that wasn't specc'd out to be anything like a server at all even ten years ago!
-
@scottalanmiller said in I can't even:
Talking to a customer who tells us that they have a two year old server "so it is time to replace it." Um, what? Who replaces two year old hardware.
Hardware...
Intel 4th Gen Desktop CPU
8GB RAM
Spinning Rust
Windows 10OMG, it's a decade old desktop that wasn't specc'd out to be anything like a server at all even ten years ago!
They are calling that a server? I can assume they are misusing the MS license then too.
-
@Obsolesce said in I can't even:
@scottalanmiller said in I can't even:
Talking to a customer who tells us that they have a two year old server "so it is time to replace it." Um, what? Who replaces two year old hardware.
Hardware...
Intel 4th Gen Desktop CPU
8GB RAM
Spinning Rust
Windows 10OMG, it's a decade old desktop that wasn't specc'd out to be anything like a server at all even ten years ago!
They are calling that a server? I can assume they are misusing the MS license then too.
It's in Africa, so licenses aren't considered legal. Not that Africa is a single place, but the treatment of copyright law is pretty standard.
-
Good backups are just wasted money and not actually important, right?
Veeam is being used for backup - awesome! Veeam is good stuff. Slight issue is that it is running on a VM, to backup VMs on the same host, itself included. Not ideal, but with a good configuration backup, it will work. If something happens to the VM, install Veeam on the replacement, restore from the latest config backup, make adjustments if necessary, and move on.
Except...
- This same VM also runs other services that aren't related to backup functions
- The backups are stored on the same physical host
- ...on the same internal host datastore
- ...on a secondary virtual disk
- ...attached to the same VM
At least there is a backup copy job to another storage target, right?
- Target is a cheap Synology DS620slim at another site
- Storage pool is 6 x 2TB RAID 5 with spinning rust
- Frequently takes 2+ days for job to succeed, but fails more frequently than it succeeds
- Job omits the largest VM (file server), probably because someone said "the other file servers at the other locations are enough for the backup because of the DFS configuration"
What was the point in spending the money on Veeam, only to configure things in a way that ensures that you don't have a solid backup if something happens to the host or the virtual disk storing the backups?!
-
@srsmith said in I can't even:
The backups are stored on the same physical host
To assist in these kinds of conversations with customers (or internal IT), I never allow people to call these backups. I say "whoa, there's no backup here, that's a file copy and nothing more, backups have a definition and this doesn't in any way meet it. We can't say we have a backup as at no point do we back anything up."
Not allowing people to say "backup" helps. As long as they are allowed to say the word, they will convince themselves and others that they are protected.
And trust me, when government agencies with backup requirements come to check, you don't want to have been using backup sloppily to mean something other than a true backup. I've had teams threatened with actual jail time for playing fast and loose with their use of the term when the SEC almost found out that they had intentionally avoided backups.
-
This thread, https://access.redhat.com/discussions/669573, likely requires a Red Hat account to view, but it wins the necro post award.
The last relevant comment was in October 2011. Here's the latest comment nearly 12 years later
-
@EddieJennings said in I can't even:
This thread, https://access.redhat.com/discussions/669573, likely requires a Red Hat account to view, but it wins the necro post award.
The last relevant comment was in October 2011. Here's the latest comment nearly 12 years later
You should log in and post this:
