Burned by Eschewing Best Practices
-
@DustinB3403 said:
RAID 5 Spinning Rust - 2 Disk Failure, without a mention of a backup device.
http://community.spiceworks.com/topic/1330485-raid-5-disk-failure?source=homepage-feed
- You are bashing a more than 10 year old system that was built when RAID5 was the standard in SMB (no matter what @scottalanmiller says about how long it has been a bad idea). You do not rebuild working equipment just because something is better. You have to consider the cost to the business and often letting it run until failure is the better option.
- He clearly stated is was a Buffalo TeraStation. If this was in a business, one can assume it was a backup target and not primary storage.
-
@JaredBusch said:
- He clearly stated is was a Buffalo TeraStation. If this was in a business, one can assume it was a backup target and not primary storage.
I'd argue that that doesn't hold up on SW. TeraStations are more that than some, but the use of consumer and "SMB" class storage like this as primary storage for running apps is discussed continuously. Maybe backup is still more common, but you definitely cannot assume that it is for backups. It is nearly daily that we talk to people that are using storage exactly like this as their primary VM stores.
-
I won't touch a Buffalo Terastation unless somebody gave me one... and even then, I'd only touch it just enough to take pictures of it and sell it on eBay... I have had horrible luck with those -- and that was just as backup targets.
-
@dafyre said:
I won't touch a Buffalo Terastation unless somebody gave me one... and even then, I'd only touch it just enough to take pictures of it and sell it on eBay... I have had horrible luck with those -- and that was just as backup targets.
I have always had good luck with them. I prefer Synology today though.
-
@JaredBusch said:
- You are bashing a more than 10 year old system that was built when RAID5 was the standard in SMB (no matter what @scottalanmiller says about how long it has been a bad idea). You do not rebuild working equipment just because something is better.
If I were bashing it (and I'm well aware of how old the equipment is) I'd be here saying "Look at this fool".
Instead I posted the topic as found on SW, without a response to the SW topic at all. Nor did I bash here. I provided what was setup 10 years ago.
The age of the device by it's self should have been cause for alarm 6 years ago, and been replaced then, in RAID10.
-
@JaredBusch said:
- You are bashing a more than 10 year old system that was built when RAID5 was the standard in SMB (no matter what @scottalanmiller says about how long it has been a bad idea). You do not rebuild working equipment just because something is better. You have to consider the cost to the business and often letting it run until failure is the better option.
Six years, I always state 2009 as the inflection point. Of course it was increasingly dangerous over time, it wasn't a "sudden" change. But 2009 is generally accepted to be the point where it just didn't make sense any longer.
It's also incredibly important to understand that just because something is standard and that lots of people do it does not suggest that it is good or that they should not have known better. In fact, it suggests the real possibility that they didn't do their due diligence and just followed the crowd hoping that things would just work out without them doing their "IT" bits.
We just had another thread talking about this - the most common thing in business, or IT or just anything is to not do it well. So if someone is only coming up to "what most do" doesn't mean that it should be taken as an acceptable level.
I've not read the thread and am not commenting on any specifics. Just that using "it was standard" as a reason to not feel that it was a bad choice isn't a good bar to set. It is our job in IT to understand and make good recommendations. If people just want to do what "everyone else is doing" they can do that without IT people to advise them.
-
@scottalanmiller said:
I've not read the thread and am not commenting on any specifics. Just that using "it was standard" as a reason to not feel that it was a bad choice isn't a good bar to set. It is our job in IT to understand and make good recommendations. If people just want to do what "everyone else is doing" they can do that without IT people to advise them.
I am most certainly not saying that it was done well, and likely the point for switching the hardware versus the cost for doing so tipped a while back for something like this in truth. Much more information would need to be know to be certain.
-
@dafyre said:
I won't touch a Buffalo Terastation unless somebody gave me one... and even then, I'd only touch it just enough to take pictures of it and sell it on eBay... I have had horrible luck with those -- and that was just as backup targets.
I've supported a few and have had to deal with Buffalo support which was very good. The devices seem to be "middling", not great but not bad. Support was quite impressive when I had to use them, though.
-
I agree that if stuff is working, often it should be kept. Not forever, things wear out and get riskier, but we often jump to replacing early in IT because it's more fun to have new stuff.
-
@JaredBusch said:
@dafyre said:
I won't touch a Buffalo Terastation unless somebody gave me one... and even then, I'd only touch it just enough to take pictures of it and sell it on eBay... I have had horrible luck with those -- and that was just as backup targets.
I have always had good luck with them. I prefer Synology today though.
Synology and ReadyNAS are really just "exceptionally good" I think. Buffalo is certainly a viable contender.
-
@DustinB3403 said:
@JaredBusch said:
- You are bashing a more than 10 year old system that was built when RAID5 was the standard in SMB (no matter what @scottalanmiller says about how long it has been a bad idea). You do not rebuild working equipment just because something is better.
If I were bashing it (and I'm well aware of how old the equipment is) I'd be here saying "Look at this fool".
Instead I posted the topic as found on SW, without a response to the SW topic at all. Nor did I bash here. I provided what was setup 10 years ago.
The age of the device by it's self should have been cause for alarm 6 years ago, and been replaced then, in RAID10.
This entire thread is nothing but a collection of posts linking to (mostly) SW highlighting mistakes. That is nothing but a thread of bashing.
As for the age of the device. Replacing something that is working just because it is old? Not a chance. As I stated in my reply to Scott, yes, likely it should have been redone as RAID10 by now, but that does not mean replacing the hardware just because.
I am working from a 6 year old desktop right now. Should I replace it?
-
@JaredBusch said:
I am working from a 6 year old desktop right now. Should I replace it?
Not necessarily... Just keep a good backup and spare drive on hand, lol.
-
@JaredBusch said:
I am working from a 6 year old desktop right now. Should I replace it?
Do you make money from that 6 year old desktop? Does that desktop hold critical data or backup data?
If either, yes replace it.
Otherwise shut it. Just because the bulk of the topics here are linked to SW has no bearing at all. If a topic like this was posted here, I'd create a link for it and say the same exact thing.
-
@DustinB3403 said:
@JaredBusch said:
I am working from a 6 year old desktop right now. Should I replace it?
Do you make money from that 6 year old desktop? Does that desktop hold critical data or backup data?
If either, yes replace it.
Otherwise shut it. Just because the bulk of the topics here are linked to SW has no bearing at all. If a topic like this was posted here, I'd create a link for it and say the same exact thing.
And that would also be bashing simply because it is in this thread.
And yes my desktop is critical to my ability to earn a paycheck. But I have no intention of replacing it anytime soon. I expect more than that out of my hardware.
-
So posting a link to SW (or any other forum) would be bashing, just because it was posted.
Wow.
You have a very skewed view of the world.
-
Be nice boys....
-
People learn from reading and doing, if there aren't examples of where something poorly setup / designed has failed, when there are safer and better ways of doing things; How'd you expect people to learn?
This is a learning topic. A topic that hopefully with SEO will help others see that maybe their plans need to be rethought as they introduce a lot of risk.
In the SW topic, the OP doesn't understand how RAID works, as he asks if replacing the failed drives will rebuild the array. So what would you say to that?
-
So this topic is on SW, and he's looking for an OpenSource Alternative to Exchange (and specifically doesn't like Office365) Nothing wrong with that.
What I do see as insane is the insistence that Cloud providers are bad. That if something happens, it's difficult to prove who's at fault. Or that it's difficult to implement (migrate) to O365.
Now @JaredBusch will probably rip on me for mentioning the topic, but the entire argument that the OP has had (brought up by others) and responded to by many pretty much sums up to misinformation and understanding.
The OP is still in the mindset that "I must run it locally to be better protected while saving money". Which needs to be addressed, but I have no better way to do it than by saying "You're insane OP"
Which it's not a best practice to use a cloud service in every instance, but in terms of uptime, cloud providers literally have to be up 99.9% of the time, or their clients will leave.
So by default, doesn't that make it a best practice?
-
Reading that guys post - just damn!
Although on the reliability stand point - I've had an inhouse Exchange server for 5+ years now. I've not had one Exchange outage (finds wood to knock on). Sure I've had ISP outages, and power outages, but never an Exchange outage.
With that in mind, MS had had countless O365 outages in the past year alone. I realize they are typically regional and often short lived, but still they are outages.
It's things like that that people sit back and say - why would I move to the cloud. MS is clearly not keeping their platform as stable as mine has been.
I understand the stats of if all, but local personal experience is really hard to overcome. Hell just look at the other recent posts around here where the guy has a two node HA setup with a NAS. His (and countless others) experience shows that his solution worked, was viable.
Frankly, as I type this I wonder if from a dollars and cents perspective if it isn't really an acceptable thing to consider. Of course you have to consider it fully (backups, recovery time, hardware replacement time, etc).
-
@DustinB3403 said:
So by default, doesn't that make it a best practice?
Not exactly. It makes it the de facto starting point, majority practice or the common reference implementation; not exactly a best practice if you want to be technical.
His approach and logic are lacking in best practices - he sounds (I did not read the post) like he is not following best practices of using real information or using logic and information to drive his decisions but instead making business decisions based on emotion. That is definitely not a best practice.
But going to hosted for anything isn't a best practice, exactly. Even if it is 99.9% of the use cases. A best practice should be something that is truly best, not really something that is "best for most people."