|
We are now on day three of a fairly large provider having their London network hosed by packet loss and the clownshoes outfit are unable to fix it or even elaborate on what they are doing. Presumably this is the old Venus network that M247 (ex-Metronet) took over. Incredibly poo poo provider.
|
# ? Sep 19, 2019 13:32 |
|
|
# ? Apr 19, 2024 08:42 |
|
"whoa whoa whoa, there's plenty of meat left in that taco!"
|
# ? Sep 19, 2019 14:28 |
|
Ghostnuke posted:Yeah. I can just go to the store and buy whatever I want, I don't see what's exciting about a vending machine. Our vending machines are subsidized, so every beverage in them, from sodas to big dumb cans of energy drinks, are 25c.
|
# ? Sep 19, 2019 17:50 |
|
Bob Morales posted:All I know is I'm never further than 150 feet from a choco taco You hiring? I have not had one of those in years....
|
# ? Sep 19, 2019 18:37 |
|
Potato Salad posted:Oracle? I kind of wish. At least then I'd probably be making stacks. I work for one of many nearly identical MDM solution providers.
|
# ? Sep 19, 2019 20:28 |
|
Bob Morales posted:"They" are going to do an experiment by removing one cpu from a server and "seeing if it runs any faster" I've had to do this with a server that ran an application that wasn't NUMA-aware, and you couldn't disable NUMA when all four CPU sockets were populated. I don't think this is gonna be the case at your company.
|
# ? Sep 20, 2019 07:42 |
|
Ugh loving Comcast. Comcast has been rocksolid for months at one location. And they are the only realistic choice there. (unless you consider 3Mbps Frontier DSL a valid choice). So I sign a contract to get better pricing with them. Weeks after I sign the contract the service turns to poo poo. They have had 4 major outages in a month. Once was the entire business day. And of course a new fiber ISP pops up offering service right after I sign that contract. Comcast wants $$$$ if the contract is broken. GGGRRRRRRRRR!
|
# ? Sep 20, 2019 14:44 |
|
For the love of loving anyone, please do not strictly use numbers as server names Server 44 is down. WTF does that do? It's been that way for years. I doubt it'll change either. Slight edit: I know I can keep a list but I'm a fan of having more descriptive names. Irritated Goat fucked around with this message at 14:53 on Sep 20, 2019 |
# ? Sep 20, 2019 14:45 |
|
Irritated Goat posted:For the love of loving anyone, please do not strictly use numbers as server names Server 44 is down. WTF does that do? Start at 69 :giggity:
|
# ? Sep 20, 2019 15:36 |
|
One of the departments here does that and they get real mad when server69 is down because they have a critical service on it. But no one but them was aware there was anything special about it because server1 through server68 are all a pile of generic batch nodes or warm spares.
|
# ? Sep 20, 2019 15:37 |
|
stevewm posted:Ugh loving Comcast. If their service is hosed can you not break contract? Might be worth running past your legal team.
|
# ? Sep 20, 2019 16:13 |
|
Thanks Ants posted:If their service is hosed can you not break contract? Might be worth running past your legal team. Read through their agreement... doesn't seem like there is anything to break out of it. Honestly just about to say gently caress it and tell them to pound sand regardless of fees, with the owners'/CEO permission of course. Edit: Fun times... Good thing we have a backup. Legendary Comcast "reliability". And yes this is the entire town going out. Not just our location. Happens at least every other week, they don't seem to care. Can't get the new provider fast enough. stevewm fucked around with this message at 20:13 on Sep 20, 2019 |
# ? Sep 20, 2019 16:28 |
|
I had a nice laugh today when I went to go figure out how to get an unresponsive Win10 machine into safe mode and it boils down to "turn if off and on again a lot until the OS finally decides to give you a prompt"
|
# ? Sep 21, 2019 01:03 |
|
Irritated Goat posted:For the love of loving anyone, please do not strictly use numbers as server names Server 44 is down. WTF does that do? Lucky for us, network stuff is all descriptive, but servers are only somewhat descriptive, except those of us not in the server teams have no idea why <site><function>345 is a domain controller and 343 is just SCCM (for example)
|
# ? Sep 21, 2019 04:32 |
|
Another person in the department brought up buying a SAN It was met with 'what happens when it goes down' I mentioned that they basically all have dual controllers, power supplies, huge arrays of disk, however, it is possible that it could fail. "So we'd have to buy two" Well....we could have two and have one mirror the other...or just back up the things that are on it. Right now the 'backup strategy' is that our servers are in two different server rooms, both in different parts of the building. So the idea is that every server has hot spare in the other server room, and if something died we could just switch the IP addresses over.
|
# ? Sep 22, 2019 17:32 |
|
Bob Morales posted:Another person in the department brought up buying a SAN I'm literally screaming at my phone 'that's not what backups are for'
|
# ? Sep 22, 2019 17:50 |
|
Bob Morales posted:Another person in the department brought up buying a SAN I know it would be extremely expensive but you could set up a dual server datacore instance to keep the two San as an active-active combo instead of hot spare(with zero downtime when one of the two goes down). Uptime seems more important than expenditure so try pushin this angle to your employer.
|
# ? Sep 22, 2019 17:53 |
|
Judge Schnoopy posted:I'm literally screaming at my phone 'that's not what backups are for' There are rando USB hard drives used for data backup too!
|
# ? Sep 22, 2019 18:00 |
|
People who fail to see the potential downsides of any new proposal in perspective to the flaming wreck that exists already are a fun breed. Yes, you're correct, Office 365 does have service outages but your email is currently served off a six year old SBS box running on a cable broadband service.
|
# ? Sep 22, 2019 18:39 |
|
Bob Morales posted:Another person in the department brought up buying a SAN I've gotten the impression that you've got a pretty great skill set. Why keep working for these small fish?
|
# ? Sep 22, 2019 18:56 |
|
Potato Salad posted:I've gotten the impression that you've got a pretty great skill set. Why keep working for these small fish? I get lured in with the old "we know we need to make improvements but we need someone to suggest them for us"
|
# ? Sep 22, 2019 19:03 |
|
Bob Morales posted:I get lured in with the old "we know we need to make improvements but we need someone to suggest them for us" I mean, it's very true in this case, they just have no intention of actually listening to the suggestions.
|
# ? Sep 22, 2019 20:26 |
Irritated Goat posted:For the love of loving anyone, please do not strictly use numbers as server names Server 44 is down. WTF does that do? Every place I’ve worked for had servers named after people, places or tv characters. Oh “Homer” is down? Great, that’s helpful. Fortunately I set up the monitoring system here so gave them descriptive names in the alerts.
|
|
# ? Sep 23, 2019 08:59 |
|
No "you killed Kenny!" messages, then?
|
# ? Sep 23, 2019 09:03 |
|
bitterandtwisted posted:Every place I’ve worked for had servers named after people, places or tv characters. Oh “Homer” is down? Great, that’s helpful. I really wanted to give our servers hip names that showed our nerdpride, but in the end I just went with SRV-DC01, SRV-SQL01 and SRV-ERP01 because I'm uncool like that.
|
# ? Sep 23, 2019 13:48 |
|
Thanks Ants posted:People who fail to see the potential downsides of any new proposal in perspective to the flaming wreck that exists already are a fun breed. Yes, you're correct, Office 365 does have service outages but your email is currently served off a six year old SBS box running on a cable broadband service. Anytime there is a brief office 365 outage of any kind the grognards come out in force to tell you how their in Prem hasn’t had an outage in 10 years. Everyone always manages to have perfect infrastructure somehow!
|
# ? Sep 23, 2019 13:49 |
|
No outages but I patch monthly and oh yeah, my server stayed up during that 72 hour blackout last February somehow. Yep! 110% uptime for me! My trusty Dell R510 and I nvented the term “zero nines(tm)” even!
|
# ? Sep 23, 2019 17:27 |
|
This isn't pissing me off so much as amusing me, mostly because I am not directly involved, just being cc'd on all the communication. We're generating reports and data to be imported into our budgeting appliance/service/cube/whatever so that department heads can do their final budget reports for 2019 and submit budget requests for 2020. The group managing the cube is getting into a tiff with one of the other groups that is generating reports with a description field isn't quoted and has comma's in it.
|
# ? Sep 23, 2019 17:58 |
|
PancakeTransmission posted:If you can't rename them, can you at least get DNS aliases made? I could slide into DNS and do it but it would be basically for me alone so I'll just keep a OneNote list and grumble about it. There's a lot of security through obscurity poo poo here so it's just 1 more thing.
|
# ? Sep 23, 2019 19:40 |
|
"So what happens when VMware starts licensing by VM?"
|
# ? Sep 23, 2019 21:33 |
|
Bob Morales posted:"So what happens when VMware starts licensing by VM?" Hey, MS switched data center from socket to cores. That was mildly annoying.
|
# ? Sep 23, 2019 22:38 |
|
sixth and maimed posted:I really wanted to give our servers hip names that showed our nerdpride, but in the end I just went with SRV-DC01, SRV-SQL01 and SRV-ERP01 because I'm uncool like that. That's what CNAMEs are for ! Pissing me off: I've got a PowerShell script that needs to collect registry values from remote machines. Since we don't have PSRemoting enabled, I have to do it the hard way with "reg query \\$target\etc." and parse the output as a string. That's only annoying me. Once I get the registry values, I use them to generate machine IDs for our backup system, and stick them into a REST call to apply a backup plan. Because my corporate masters use a self-signed certificate, the REST call has to go out over http, not https. That's only annoying me. I need to put credentials into the call, and since it's http only, I'm restricting use of that script to the actual backup server. That's only annoying me. The 'reg query \\...' call fails reliably in PowerShell, but works fine in cmd.exe. It also works in PowerShell on my laptop, the other backup server, and one of the RDP jump boxes. I eventually figure out that in cmd.exe the 32-bit version fails but the 64-bit version works. After a little more digging... cmd.exe has its PROCESSOR_ARCHITECTURE set to AMD64, but PowerShell has $Env:PROCESSOR_ARCHITECTURE set to x86. I log out and log back in, and the script works. I have no idea how PS got stuck in 32-bit mode, but that's what's pissing me off.
|
# ? Sep 23, 2019 22:50 |
|
Sickening posted:Anytime there is a brief office 365 outage of any kind the grognards come out in force to tell you how their in Prem hasn’t had an outage in 10 years. Everyone always manages to have perfect infrastructure somehow! on-prem infrastructure tends to work flawlessly until it fails catastrophically while cloud infra tends to fail periodically but it'll always come back and that is Not Your Problem(tm) So many email servers are running on dell pizza boxes out there. And yeah it works great until someone doesn't replace a failed disk because the SNMP server that was originally set up to monitor it has failed to poll it for the last 1596 intervals because someone quickly upgraded the ilo controller to fix the java console and the guy that monitors the SNMP server has been on leave
|
# ? Sep 23, 2019 23:44 |
|
gently caress the out of date version of Bootstrap on this subsite. It being out of date has added 20 or so hours to what should have been a simple project. Luckily the subsite will be updated soon.
|
# ? Sep 24, 2019 00:27 |
|
abigserve posted:on-prem infrastructure tends to work flawlessly until it fails catastrophically while cloud infra tends to fail periodically but it'll always come back and that is Not Your Problem(tm) None of those on Prem setups have better network consistency than 0365. Almost all of them don’t factor in their internet outages because being still able to email internally somehow isn’t an outage. Nobody runs a 100% clean shop that is outage proof. Nobody escapes partial outages at the very least for an infinite amount of reasons.
|
# ? Sep 24, 2019 00:57 |
|
Bob Morales posted:"So what happens when VMware starts licensing by VM?" This is a person who has worked with Microsoft licensing before. Honestly, I can't entirely blame them for their once-bitten, twice shy outlook on this issue.
|
# ? Sep 24, 2019 02:02 |
|
I was never good with probability But 18 servers instead of 3 means you're six times more likely to have a failure right?
|
# ? Sep 24, 2019 02:41 |
|
Our Magento contractor left an ancient version of adminer open to the public, and we got some malware on the site. Nobody but me sees this as a problem.
|
# ? Sep 24, 2019 08:13 |
|
Bob Morales posted:"So what happens when VMware starts licensing by VM?" VMware lets you buy some of their products on a VM license already(vSphere for desktops for instance), on low units(1-30) it's actually cheaper than per core/server licensing. https://www.vmware.com/support/support-resources/licensing/per-vm.html Beside VDI, vSphere is sold/leased by VM for hosting purposes (SPP/Cloud Provider). SlowBloke fucked around with this message at 12:35 on Sep 24, 2019 |
# ? Sep 24, 2019 12:19 |
|
|
# ? Apr 19, 2024 08:42 |
|
Let me share with you what we call "Mail drives" What's a mail drive? Well, that's when someone complains enough about Outlook being slow that we buy a drive just to run their PST file on! So, even though you have a 256GB SSD as your C: drive, your 20-30GB mail file gets moved to a: 160GB or 250GB HD we found that should be in the trash 1TB HD that we buy new I argued that in no way could this improve performance for anyone. It was meant with "MORE SPINDLES MORE POWAH" I then said but even a dogshit OEM Dell SSD = 100 spindles
|
# ? Sep 24, 2019 13:57 |