The new forums will be named Coin Return (based on the most recent vote)! You can check on the status and timeline of the transition to the new forums here.
Please vote in the Forum Structure Poll. Polling will close at 2PM EST on January 21, 2025.
I'm glad to be getting out of working in hosting, but I will be sad not to be able to read incredibly stupid support tickets from time to time.
I think it is funny that you think you'll never see a stupid support ticket again. They are a fact of life. :rotate:
Are you suggesting that while the users may change, the idiocy never does?
0
RandomHajileNot actually a SnatcherThe New KremlinRegistered Userregular
Speaking of the idiocy of users...
So we have this dumb portal thing that basically just links you to URLs of various other intranet sites. Sounds like that would be a perfect thing to have be an HTML page of links, right? Nope, it's a compiled executable that is around 5MB, which heavily depends on Flash player for some reason. (The reason is that the guy who designed it is an idiot.) We spent thousands of dollars on this thing to a parent company. As the Intranet admin, I had major objections to it in cost and in principle, and even went so far as to re-design it in HTML (and around 100KB) in a couple hours. Nope, we still went with the Flash one. They even put a link to the damn intranet on this thing. Further, they wanted us to deploy it to users' desktops and have it auto-run at login, but we were able to fight that one off.
I still duplicated the links to the things on the portal over on our intranet, because I've spent 8 years getting people to use the intranet. The intranet is linked from our network menus, and has been since 2000, before I started here.
So today I come in and have a tech at one of our remote locations complaining about it just showing a white screen. To make a long story short, Flash updated for a bunch of people and now it doesn't work for them. After a short discussion with my boss, we decided to finally kill it. The tech got another phone call that went something like this:
User: "Well, you know, there's not really any other good way to get to application site."
Tech: ".....it's on the intranet, and always has been."
User: "No, I don't think so."
Tech: "Well, it's there."
User: "But I use the portal to get to the intranet. How will I get there now?"
I purchased a new Dell server and opted for the front-loading 2.5" hard drive chassis. It supports 16 drives, of which I will probably only ever use 4.
I also purchased four 512GB Samsung 850 Pro SSDs. I'm going to run a few VMs on this thing (Server 2012 Standard R2 and Hyper-V), and I thought it might be nice to put the 4 drives in RAID10 and run the VMs off it (the OS is on its own drive). This leaves me two options:
1) Use the built in RAID controller to make the RAID (preferred).
2) Use Windows to do a software RAID.
The first option works great and Windows sees a 1TB drive, but I lose direct access to the individual drives. This means I can't update firmware or run individual SMART checks with Samsung's Magician software.
The second option is apparently not an option at all within Windows. I can create a striped volume or a mirrored volume, but I can't stripe mirrors. I can, however, access each individual drive and see their respective statuses.
The previous server which I inherited had a three-drive RAID5 setup, in which two of the drives died without any notifications. I want to make sure there's actual drive health monitoring this time and I prefer RAID10 over RAID5. I'm not sure the best way to set everything up, though.
I purchased a new Dell server and opted for the front-loading 2.5" hard drive chassis. It supports 16 drives, of which I will probably only ever use 4.
I also purchased four 512GB Samsung 850 Pro SSDs. I'm going to run a few VMs on this thing (Server 2012 Standard R2 and Hyper-V), and I thought it might be nice to put the 4 drives in RAID10 and run the VMs off it (the OS is on its own drive). This leaves me two options:
1) Use the built in RAID controller to make the RAID (preferred).
2) Use Windows to do a software RAID.
The first option works great and Windows sees a 1TB drive, but I lose direct access to the individual drives. This means I can't update firmware or run individual SMART checks with Samsung's Magician software.
The second option is apparently not an option at all within Windows. I can create a striped volume or a mirrored volume, but I can't stripe mirrors. I can, however, access each individual drive and see their respective statuses.
The previous server which I inherited had a three-drive RAID5 setup, in which two of the drives died without any notifications. I want to make sure there's actual drive health monitoring this time and I prefer RAID10 over RAID5. I'm not sure the best way to set everything up, though.
Thoughts?
I've never implemented a RAID 10, but on our VM host, I set the OS array as a RAID 1 and the other array as a 4-disk RAID 5 (striping with parity), which provided greater redundancy. All HDDs were standard with 15k RPM and 500GB. While the OS hosts Hyper-V, all vHDDs are stored on the other array. The VMs it hosts are
SharePoint server
SQL server
Windows desktop for GIS usage
Web server (soon to be new SharePoint with document management)
SQL server for document management for #4
The only issue we've run into was with memory and fucking NUMA nodes*. Basically, just spec the shit out of it for tons of memory to be allocated properly. However, none of the systems are crazy intensive (e.g., Exchange, massive SQL DBs accessed by hundreds at a time).
*I really wish there was good documentation on NUMA nodes. Everything I've read all deals with 4-core (or less) systems and how to calculate NUMA node memory allocation. But when trying to apply it to any system with more than 4 logical cores, the whole calculation makes no sense. The calculation was never designed for multiple 8-core systems, which makes the NUMA node calculation ridiculously small to the point of being laughable. It basically turns into a trial and error to determine the proper NUMA node memory allocation size. Fortunately, none of our servers seem to suffer from this to a noticeable extent because they aren't heavy servers. The more intensive of a system the VM is, the more of the impact NUMA nodes memory allocation will have.
Le_Goat on
While I agree that being insensitive is an issue, so is being oversensitive.
As I said above, I've never setup RAID 10 before, but my guess would be to either:
Setup a mirrored pair, then stripe
Setup a stripe array, then mirror it
I don't know which would be faster, but I'd assume you'd want to stripe across the mirrored pair. You'd essentially be creating two separate RAID arrays. I don't think you'd be able to set it up at the same time. I could be wrong.
Le_Goat on
While I agree that being insensitive is an issue, so is being oversensitive.
0
RandomHajileNot actually a SnatcherThe New KremlinRegistered Userregular
I purchased a new Dell server and opted for the front-loading 2.5" hard drive chassis. It supports 16 drives, of which I will probably only ever use 4.
I also purchased four 512GB Samsung 850 Pro SSDs. I'm going to run a few VMs on this thing (Server 2012 Standard R2 and Hyper-V), and I thought it might be nice to put the 4 drives in RAID10 and run the VMs off it (the OS is on its own drive). This leaves me two options:
1) Use the built in RAID controller to make the RAID (preferred).
2) Use Windows to do a software RAID.
The first option works great and Windows sees a 1TB drive, but I lose direct access to the individual drives. This means I can't update firmware or run individual SMART checks with Samsung's Magician software.
The second option is apparently not an option at all within Windows. I can create a striped volume or a mirrored volume, but I can't stripe mirrors. I can, however, access each individual drive and see their respective statuses.
The previous server which I inherited had a three-drive RAID5 setup, in which two of the drives died without any notifications. I want to make sure there's actual drive health monitoring this time and I prefer RAID10 over RAID5. I'm not sure the best way to set everything up, though.
Thoughts?
I assume it is a PERC card of some variety? All of that can be managed with OpenManage, which should come free with your server. You may have to check in on it (which can be done remotely once you install OM). The whole point of the PERC card is that it presents the array as one big drive. Otherwise you would have to setup drive passthrough, but I really wouldn't recommend Windows Software RAID if you're running VMs on it.
If it additionally came with a DRAC card (even just Express) of some sort, it should be able to be setup to email you if there is a failure.
I purchased a new Dell server and opted for the front-loading 2.5" hard drive chassis. It supports 16 drives, of which I will probably only ever use 4.
I also purchased four 512GB Samsung 850 Pro SSDs. I'm going to run a few VMs on this thing (Server 2012 Standard R2 and Hyper-V), and I thought it might be nice to put the 4 drives in RAID10 and run the VMs off it (the OS is on its own drive). This leaves me two options:
1) Use the built in RAID controller to make the RAID (preferred).
2) Use Windows to do a software RAID.
The first option works great and Windows sees a 1TB drive, but I lose direct access to the individual drives. This means I can't update firmware or run individual SMART checks with Samsung's Magician software.
The second option is apparently not an option at all within Windows. I can create a striped volume or a mirrored volume, but I can't stripe mirrors. I can, however, access each individual drive and see their respective statuses.
The previous server which I inherited had a three-drive RAID5 setup, in which two of the drives died without any notifications. I want to make sure there's actual drive health monitoring this time and I prefer RAID10 over RAID5. I'm not sure the best way to set everything up, though.
Thoughts?
I assume it is a PERC card of some variety? All of that can be managed with OpenManage, which should come free with your server. You may have to check in on it (which can be done remotely once you install OM). The whole point of the PERC card is that it presents the array as one big drive. Otherwise you would have to setup drive passthrough, but I really wouldn't recommend Windows Software RAID if you're running VMs on it.
If it additionally came with a DRAC card (even just Express) of some sort, it should be able to be setup to email you if there is a failure.
Yes, it's a PERC H330. I'll look into OpenManage, thanks!
You just found it I'm feeding you your own damn object from when you found it!
I'm fucking NT AUTHORITY you do what I goddamn say aaaaaarghghgejejhhfkldm
life's a game that you're bound to lose / like using a hammer to pound in screws
fuck up once and you break your thumb / if you're happy at all then you're god damn dumb
that's right we're on a fucked up cruise / God is dead but at least we have booze
bad things happen, no one knows why / the sun burns out and everyone dies
I've now discovered the whoami command, thanks to you, and that it's in bash, too.
I'm going to abuse that in every screenshot and doc I write from now on.
what's the point of opening CMDs as NT AUTHORITY\SYSTEM if you can't revel in your raw unlimited power
well except I'm limited in taking ownership of this god damn folder
aioua on
life's a game that you're bound to lose / like using a hammer to pound in screws
fuck up once and you break your thumb / if you're happy at all then you're god damn dumb
that's right we're on a fucked up cruise / God is dead but at least we have booze
bad things happen, no one knows why / the sun burns out and everyone dies
pretty sure at this point it's actually the storage controller biting the dust
life's a game that you're bound to lose / like using a hammer to pound in screws
fuck up once and you break your thumb / if you're happy at all then you're god damn dumb
that's right we're on a fucked up cruise / God is dead but at least we have booze
bad things happen, no one knows why / the sun burns out and everyone dies
can you quit bitching about how my system doesn't have a label yet and how you think I've configured the OS wrong or how I should have a crossover cable to set up a sql cluster (wtf?) and just SET UP MY DAMN LUNS
life's a game that you're bound to lose / like using a hammer to pound in screws
fuck up once and you break your thumb / if you're happy at all then you're god damn dumb
that's right we're on a fucked up cruise / God is dead but at least we have booze
bad things happen, no one knows why / the sun burns out and everyone dies
This is the SAN team that has has multiple data losses in the six months I've worked for this team.
And not like, from some catastrophic failure, but from them not replacing disks.
I'm not surprised this dude's networking knowledge is stuck in 1992.
life's a game that you're bound to lose / like using a hammer to pound in screws
fuck up once and you break your thumb / if you're happy at all then you're god damn dumb
that's right we're on a fucked up cruise / God is dead but at least we have booze
bad things happen, no one knows why / the sun burns out and everyone dies
Are they just leaving dead disks in the chassis for like months?
yes
no one has even been fired for this
MEGACORP is a real stupid company and I'll be so happy to leave.
In other news I fixed my mystery there/not there files.
Turned it off and on again.
life's a game that you're bound to lose / like using a hammer to pound in screws
fuck up once and you break your thumb / if you're happy at all then you're god damn dumb
that's right we're on a fucked up cruise / God is dead but at least we have booze
bad things happen, no one knows why / the sun burns out and everyone dies
can you quit bitching about how my system doesn't have a label yet and how you think I've configured the OS wrong or how I should have a crossover cable to set up a sql cluster (wtf?) and just SET UP MY DAMN LUNS
Who the hell is this guy? I mean, I am, as one of my jobs, the SAN administrator. When we got the thing, I had no idea what I was doing, nor did I understand how the networking was setup. We even had a stupid major failure (but didn't lose any data) due to not replacing disks and having a fundamental misunderstanding of how the RAID was actually setup (vs. how it appeared to be setup in the GUI). But I learned, and I set up our new SAN box entirely with no outside help. I'm no expert, but I'm pretty proud of where I am with that 10% of my job.
And I can't think of any good reason why you'd need a crossover cable for anything when setting up LUNs from a server. The only bad reason I can think of is that the SAN box doesn't have it's own dedicated switch, so they're just hooking servers up to the SAN directly? That would be reeeeeeaallly dumb, but I guess it would work.
can you quit bitching about how my system doesn't have a label yet and how you think I've configured the OS wrong or how I should have a crossover cable to set up a sql cluster (wtf?) and just SET UP MY DAMN LUNS
Who the hell is this guy? I mean, I am, as one of my jobs, the SAN administrator. When we got the thing, I had no idea what I was doing, nor did I understand how the networking was setup. We even had a stupid major failure (but didn't lose any data) due to not replacing disks and having a fundamental misunderstanding of how the RAID was actually setup (vs. how it appeared to be setup in the GUI). But I learned, and I set up our new SAN box entirely with no outside help. I'm no expert, but I'm pretty proud of where I am with that 10% of my job.
And I can't think of any good reason why you'd need a crossover cable for anything when setting up LUNs from a server. The only bad reason I can think of is that the SAN box doesn't have it's own dedicated switch, so they're just hooking servers up to the SAN directly? That would be reeeeeeaallly dumb, but I guess it would work.
I've come to the conclusion that 98% of the people on this planet do not have a functioning brain and don't understand logic or sequence of events (this happens, so that happens).
And it makes me realize that someone who can take something, break it down into pieces, and do work with it, is worth their weight in fucking gold. The rest of the planet capitalizes on that 2% that can actually think critically in their day to day work.
These people don't even have to be high up the chain, they're probably not, they're probably at the lowest, and the rest excel by putting their boot on the other's face and using them as a stepping stone.
the private network thing has nothing to do with the SAN, the SAN dude is just sticking his head in where it doesn't belong and offering us sysadmin advice (and/or possibly trying to make us look bad and/or looking for flimsy excuses to not do his work)
ofc my boss comes in and is all "yeah of course we do that" never mind that none of my coworkers have heard of it before
Still it seems like having a private network between your cluster nodes is of fairly limited utility. Like, ok I guess it can more gracefully failover when one of the nodes loses corp networking (but not the private network).
aioua on
life's a game that you're bound to lose / like using a hammer to pound in screws
fuck up once and you break your thumb / if you're happy at all then you're god damn dumb
that's right we're on a fucked up cruise / God is dead but at least we have booze
bad things happen, no one knows why / the sun burns out and everyone dies
don't most network cards/switches/hubs auto sense what type of cable it is and switch to the appropriate one anyways? I don't think I've needed an x-over cable for nearly 20 years.
don't most network cards/switches/hubs auto sense what type of cable it is and switch to the appropriate one anyways? I don't think I've needed an x-over cable for nearly 20 years.
yes, that bit was pure anachronism
but my boss wants the machines connected directly to each other over one NIC, and to the corp network over the other NIC
life's a game that you're bound to lose / like using a hammer to pound in screws
fuck up once and you break your thumb / if you're happy at all then you're god damn dumb
that's right we're on a fucked up cruise / God is dead but at least we have booze
bad things happen, no one knows why / the sun burns out and everyone dies
don't most network cards/switches/hubs auto sense what type of cable it is and switch to the appropriate one anyways? I don't think I've needed an x-over cable for nearly 20 years.
yes, that bit was pure anachronism
but my boss wants the machines connected directly to each other over one NIC, and to the corp network over the other NIC
Are they acting as a ring network or something? Does that even work?
0
RandomHajileNot actually a SnatcherThe New KremlinRegistered Userregular
don't most network cards/switches/hubs auto sense what type of cable it is and switch to the appropriate one anyways? I don't think I've needed an x-over cable for nearly 20 years.
I haven't needed a crossover cable for years, either, but that's because we have dedicated switches like a real big-boy company.
don't most network cards/switches/hubs auto sense what type of cable it is and switch to the appropriate one anyways? I don't think I've needed an x-over cable for nearly 20 years.
yes, that bit was pure anachronism
but my boss wants the machines connected directly to each other over one NIC, and to the corp network over the other NIC
Are they acting as a ring network or something? Does that even work?
point to point, just manually set the IPs.
supposedly Failover Manager will see the connection and grab it and use it for whatever it is that it does
life's a game that you're bound to lose / like using a hammer to pound in screws
fuck up once and you break your thumb / if you're happy at all then you're god damn dumb
that's right we're on a fucked up cruise / God is dead but at least we have booze
bad things happen, no one knows why / the sun burns out and everyone dies
can you quit bitching about how my system doesn't have a label yet and how you think I've configured the OS wrong or how I should have a crossover cable to set up a sql cluster (wtf?) and just SET UP MY DAMN LUNS
Who the hell is this guy? I mean, I am, as one of my jobs, the SAN administrator. When we got the thing, I had no idea what I was doing, nor did I understand how the networking was setup. We even had a stupid major failure (but didn't lose any data) due to not replacing disks and having a fundamental misunderstanding of how the RAID was actually setup (vs. how it appeared to be setup in the GUI). But I learned, and I set up our new SAN box entirely with no outside help. I'm no expert, but I'm pretty proud of where I am with that 10% of my job.
And I can't think of any good reason why you'd need a crossover cable for anything when setting up LUNs from a server. The only bad reason I can think of is that the SAN box doesn't have it's own dedicated switch, so they're just hooking servers up to the SAN directly? That would be reeeeeeaallly dumb, but I guess it would work.
I've come to the conclusion that 98% of the people on this planet do not have a functioning brain and don't understand logic or sequence of events (this happens, so that happens).
And it makes me realize that someone who can take something, break it down into pieces, and do work with it, is worth their weight in fucking gold. The rest of the planet capitalizes on that 2% that can actually think critically in their day to day work.
These people don't even have to be high up the chain, they're probably not, they're probably at the lowest, and the rest excel by putting their boot on the other's face and using them as a stepping stone.
To be honest, you basically just described troubleshooting. A lot of people don't get how to do that, which is why we get paid to do shit. The logic involved is some of the most basic and fundamental types, like boolean logic. This day and age, there are so many things that are done automatically for people because of a computing device. With that comes being lazy and forgetting how to think that way. If you can problem solve, there are good jobs out there for you.
What's funny is that you can substitute other skills in there and your statement will still make sense.
While I agree that being insensitive is an issue, so is being oversensitive.
still you don't need us to turn on a computer or plug it in
We're there to replace hard drives in a SAN.
until they invent a robot to do that for us
I mean I guess you could just have lots of hot spares but how cool would it be to have an autonomous bot that looks for amber lights and then swaps drives with a robot arm
life's a game that you're bound to lose / like using a hammer to pound in screws
fuck up once and you break your thumb / if you're happy at all then you're god damn dumb
that's right we're on a fucked up cruise / God is dead but at least we have booze
bad things happen, no one knows why / the sun burns out and everyone dies
still you don't need us to turn on a computer or plug it in
We're there to replace hard drives in a SAN.
until they invent a robot to do that for us
I mean I guess you could just have lots of hot spares but how cool would it be to have an autonomous bot that looks for amber lights and then swaps drives with a robot arm
Bad HDDs trigger an event in the logs; event id 7, to be exact. Theoretically, you have a bot setup to trigger when event ID 7 hits, and then the drive is hot swapped out, allowing the RAID to build on its own without human intervention.
Of course, after this comes Skynet and then the end of mankind as we know it.
Le_Goat on
While I agree that being insensitive is an issue, so is being oversensitive.
Posts
I think it is funny that you think you'll never see a stupid support ticket again. They are a fact of life. :rotate:
Maybe IT should be notified? You think?!? Gah.
Are you suggesting that while the users may change, the idiocy never does?
So we have this dumb portal thing that basically just links you to URLs of various other intranet sites. Sounds like that would be a perfect thing to have be an HTML page of links, right? Nope, it's a compiled executable that is around 5MB, which heavily depends on Flash player for some reason. (The reason is that the guy who designed it is an idiot.) We spent thousands of dollars on this thing to a parent company. As the Intranet admin, I had major objections to it in cost and in principle, and even went so far as to re-design it in HTML (and around 100KB) in a couple hours. Nope, we still went with the Flash one. They even put a link to the damn intranet on this thing. Further, they wanted us to deploy it to users' desktops and have it auto-run at login, but we were able to fight that one off.
I still duplicated the links to the things on the portal over on our intranet, because I've spent 8 years getting people to use the intranet. The intranet is linked from our network menus, and has been since 2000, before I started here.
So today I come in and have a tech at one of our remote locations complaining about it just showing a white screen. To make a long story short, Flash updated for a bunch of people and now it doesn't work for them. After a short discussion with my boss, we decided to finally kill it. The tech got another phone call that went something like this:
User: "Well, you know, there's not really any other good way to get to application site."
Tech: ".....it's on the intranet, and always has been."
User: "No, I don't think so."
Tech: "Well, it's there."
User: "But I use the portal to get to the intranet. How will I get there now?"
Ugh.
This is a clickable link to my Steam Profile.
I also purchased four 512GB Samsung 850 Pro SSDs. I'm going to run a few VMs on this thing (Server 2012 Standard R2 and Hyper-V), and I thought it might be nice to put the 4 drives in RAID10 and run the VMs off it (the OS is on its own drive). This leaves me two options:
1) Use the built in RAID controller to make the RAID (preferred).
2) Use Windows to do a software RAID.
The first option works great and Windows sees a 1TB drive, but I lose direct access to the individual drives. This means I can't update firmware or run individual SMART checks with Samsung's Magician software.
The second option is apparently not an option at all within Windows. I can create a striped volume or a mirrored volume, but I can't stripe mirrors. I can, however, access each individual drive and see their respective statuses.
The previous server which I inherited had a three-drive RAID5 setup, in which two of the drives died without any notifications. I want to make sure there's actual drive health monitoring this time and I prefer RAID10 over RAID5. I'm not sure the best way to set everything up, though.
Thoughts?
The only issue we've run into was with memory and fucking NUMA nodes*. Basically, just spec the shit out of it for tons of memory to be allocated properly. However, none of the systems are crazy intensive (e.g., Exchange, massive SQL DBs accessed by hundreds at a time).
*I really wish there was good documentation on NUMA nodes. Everything I've read all deals with 4-core (or less) systems and how to calculate NUMA node memory allocation. But when trying to apply it to any system with more than 4 logical cores, the whole calculation makes no sense. The calculation was never designed for multiple 8-core systems, which makes the NUMA node calculation ridiculously small to the point of being laughable. It basically turns into a trial and error to determine the proper NUMA node memory allocation size. Fortunately, none of our servers seem to suffer from this to a noticeable extent because they aren't heavy servers. The more intensive of a system the VM is, the more of the impact NUMA nodes memory allocation will have.
As I said above, I've never setup RAID 10 before, but my guess would be to either:
I don't know which would be faster, but I'd assume you'd want to stripe across the mirrored pair. You'd essentially be creating two separate RAID arrays. I don't think you'd be able to set it up at the same time. I could be wrong.
If it additionally came with a DRAC card (even just Express) of some sort, it should be able to be setup to email you if there is a failure.
This is a clickable link to my Steam Profile.
Yes, it's a PERC H330. I'll look into OpenManage, thanks!
What do you fucking mean you can't find it?
You just found it I'm feeding you your own damn object from when you found it!
I'm fucking NT AUTHORITY you do what I goddamn say aaaaaarghghgejejhhfkldm
fuck up once and you break your thumb / if you're happy at all then you're god damn dumb
that's right we're on a fucked up cruise / God is dead but at least we have booze
bad things happen, no one knows why / the sun burns out and everyone dies
I'm going to abuse that in every screenshot and doc I write from now on.
what's the point of opening CMDs as NT AUTHORITY\SYSTEM if you can't revel in your raw unlimited power
well except I'm limited in taking ownership of this god damn folder
fuck up once and you break your thumb / if you're happy at all then you're god damn dumb
that's right we're on a fucked up cruise / God is dead but at least we have booze
bad things happen, no one knows why / the sun burns out and everyone dies
naw it can take the object:
pretty sure at this point it's actually the storage controller biting the dust
fuck up once and you break your thumb / if you're happy at all then you're god damn dumb
that's right we're on a fucked up cruise / God is dead but at least we have booze
bad things happen, no one knows why / the sun burns out and everyone dies
can you quit bitching about how my system doesn't have a label yet and how you think I've configured the OS wrong or how I should have a crossover cable to set up a sql cluster (wtf?) and just SET UP MY DAMN LUNS
fuck up once and you break your thumb / if you're happy at all then you're god damn dumb
that's right we're on a fucked up cruise / God is dead but at least we have booze
bad things happen, no one knows why / the sun burns out and everyone dies
Wait.
What.
And not like, from some catastrophic failure, but from them not replacing disks.
I'm not surprised this dude's networking knowledge is stuck in 1992.
fuck up once and you break your thumb / if you're happy at all then you're god damn dumb
that's right we're on a fucked up cruise / God is dead but at least we have booze
bad things happen, no one knows why / the sun burns out and everyone dies
What..
That's their only job?
How.
Are they just leaving dead disks in the chassis for like months?
yes
no one has even been fired for this
MEGACORP is a real stupid company and I'll be so happy to leave.
In other news I fixed my mystery there/not there files.
fuck up once and you break your thumb / if you're happy at all then you're god damn dumb
that's right we're on a fucked up cruise / God is dead but at least we have booze
bad things happen, no one knows why / the sun burns out and everyone dies
And I can't think of any good reason why you'd need a crossover cable for anything when setting up LUNs from a server. The only bad reason I can think of is that the SAN box doesn't have it's own dedicated switch, so they're just hooking servers up to the SAN directly? That would be reeeeeeaallly dumb, but I guess it would work.
This is a clickable link to my Steam Profile.
I've come to the conclusion that 98% of the people on this planet do not have a functioning brain and don't understand logic or sequence of events (this happens, so that happens).
And it makes me realize that someone who can take something, break it down into pieces, and do work with it, is worth their weight in fucking gold. The rest of the planet capitalizes on that 2% that can actually think critically in their day to day work.
These people don't even have to be high up the chain, they're probably not, they're probably at the lowest, and the rest excel by putting their boot on the other's face and using them as a stepping stone.
Most people would spend money to fix it rather than try to fix it themselves.
ofc my boss comes in and is all "yeah of course we do that" never mind that none of my coworkers have heard of it before
Still it seems like having a private network between your cluster nodes is of fairly limited utility. Like, ok I guess it can more gracefully failover when one of the nodes loses corp networking (but not the private network).
fuck up once and you break your thumb / if you're happy at all then you're god damn dumb
that's right we're on a fucked up cruise / God is dead but at least we have booze
bad things happen, no one knows why / the sun burns out and everyone dies
yes, that bit was pure anachronism
but my boss wants the machines connected directly to each other over one NIC, and to the corp network over the other NIC
fuck up once and you break your thumb / if you're happy at all then you're god damn dumb
that's right we're on a fucked up cruise / God is dead but at least we have booze
bad things happen, no one knows why / the sun burns out and everyone dies
Are they acting as a ring network or something? Does that even work?
But yes, a lot of ports do autosense.
This is a clickable link to my Steam Profile.
point to point, just manually set the IPs.
supposedly Failover Manager will see the connection and grab it and use it for whatever it is that it does
fuck up once and you break your thumb / if you're happy at all then you're god damn dumb
that's right we're on a fucked up cruise / God is dead but at least we have booze
bad things happen, no one knows why / the sun burns out and everyone dies
What's funny is that you can substitute other skills in there and your statement will still make sense.
We're there to replace hard drives in a SAN.
until they invent a robot to do that for us
I mean I guess you could just have lots of hot spares but how cool would it be to have an autonomous bot that looks for amber lights and then swaps drives with a robot arm
fuck up once and you break your thumb / if you're happy at all then you're god damn dumb
that's right we're on a fucked up cruise / God is dead but at least we have booze
bad things happen, no one knows why / the sun burns out and everyone dies
Of course, after this comes Skynet and then the end of mankind as we know it.