why are most network/server issues fixed with a reboot???

healthyboy

how come a lot of server and network issues r fixed with a reboot??

In your role when a network or a issue comes wats da percentage of them being fixed by a reboot in your role?

Jasiono

In the production network at my job a reboot is the ultimate last resort. While there are a lot of redundancies you still have to take into effect the convergence time, depending on how many you are rebooting.

healthyboy

yeah ofcourse, thats like everywhere else but still all it needs is a reboot just like how u work on end users

ptilsen

I would actually say that relatively few issues are solved by a reboot, with even fewer requiring a reboot. But there are many technical reasons, depending on the equipment or operating system, that a reboot can clear problems.

A reboot restarts all services; a cold boot will clear RAM, and even a reboot will mostly clear memory; a reboot will disconnect everything.

There are many more reasons, but those are just some big, obvious ones. That said, "reboot and pray," as I like to call it, should almost never be your first resort, and should usually be your last resort. Most software and network problems can be isolated to a specific service or device with good troubleshooting skills. All one needs is a good understanding of networking and the OSI model, some familiarity with the software in question, and well rounded OS and hardware knowledge.

joshmadakor

Pretty much sums up my thoughts.

cxzar20

Most networking issues are not solved from a reboot, IOS doesn't work like Windows. A reboot only happens if there is a memory leak issue causing CPU to rise or something like that (which will later require an IOS upgrade). Nobody is going to touch my network infrastructure if they go around rebooting things. In the networking world there is almost always an RFC reason why something isn't working.

blargoe

Rebooting is not a solution. Yes, it often clears an error condition or poor system performance, but just relieves the symptoms for a little while. If your protocol for providing support consists of rebooting first, unfortunately you will not get very far in your IT career.

noobsrevenge

Really? And to think people take years and years to learn troubleshooting skills when they could just reboot.

Akaricloud

Rebooting should be one of the last options for servers, not the first. For client computers it's not a bad option but in an environment where hundreds or thousands of people depend on that server, it's not a good option.

I mainly see people do it to restart the services in the correct order since it can be difficult to do without knowing which depend on eachother. I usually try to script these in advance so if I do need to restart services for one reason or another, then I don't need to waste time rebooting.

it_consultant

cxzar20 wrote: »

Most networking issues are not solved from a reboot, IOS doesn't work like Windows. A reboot only happens if there is a memory leak issue causing CPU to rise or something like that (which will later require an IOS upgrade). Nobody is going to touch my network infrastructure if they go around rebooting things. In the networking world there is almost always an RFC reason why something isn't working.

Oh yeah, I have NEVER had to punch a cisco router in the eye after it locked up. Although generally speaking, since routers and switches are such simple devices compared to servers they don't require a hard punch nearly as often.

Have you ever talked to Dell or HP laptop support? One of the first troubleshooting steps they give you is to unplug the thing, pull the battery out, wait a few seconds, plug the battery back in (or not), plug it into AC and power it up.

Below assembly and machine code you still have circuits which are controlled by microscopic transistors which are turned off and on (representing binary one and zero) by tiny amounts of voltage. It is less common now, but those circuits get hot, they fail to operate properly, etc etc. That is on top of whatever software environment which might be going through sub-routines that are not obvious to you, which are causing issues.

NetworkVeteran

healthyboy wrote: »

how come a lot of server and network issues r fixed with a reboot??

1) The longer a piece of software is running, the more likely it is to produce unusual or erroneous behavior, and the effects of those behaviors accumulate over time. A reboot sets the state of the software back to an initial, known, clean state.

2) A person who lacks domain troubleshooting skills to quickly ID and understand unusual behaviors will often turn to rebooting and that is the "solution". Since they neither understood nor targetted the actual issue, the behavior will surface again and again, being repeatedly "fixed" by rebooting.

NetworkVeteran

healthyboy wrote: »

In your role when a network or a issue comes wats da percentage of them being fixed by a reboot in your role?

Less than 1%? On my home network, it's closer to 10%. Low SLA requirements.

the_Grinch

I try to avoid reboots when I can, but often there are times when you tried everything else. What always struck me was when a reboot was required, but you are never informed of that fact. I had a doctor who got a widescreen monitor and had turned it to show patients MRI's. He called because it wasn't displaying properly. I did all the normal troubleshooting and after looking all over found a drive for the monitor. It wasn't a standard Dell drive, had to extract it myself and install via a dll (I believe). After that, the thing still would display properly. He had to run and since it was a personal machine he was using said he'd give Dell a call and then return the monitor if that was the issue. It just so happens he calls me back and says "what is the one thing you guys always say to do that we didn't do?" I replied "please don't say reboot" and he just laughed himself silly for about 10 minutes. That was the first thing the Dell tech said when he explained what we did and the monitor displayed properly instantly. So there is a time and place, plus the "reboot and pray".

kurosaki00

because its windows?
no but really, usually is some service that had issues, it can snowball into a full system issue
usually a service restart will suffix, if it continues then it becomes a problem
but many times yeah a restart of service will fix stuff

On networks... I dont know
I dont see much problems in networking that can be fixed with a reboot

bigdogz

I think this used to happen back in the old days of hub, dumb switches, Workgroups for Windows, OS/2 and NT 3.0 servers. It did not happen as much with Novell and Unix servers because restarting the services needed was handled better with those systems at the time. If you did that to a Windows server it did not really like it. Back in those days there were a great deal of memory leaks on all types of firmware. I have also had this happen to me on a SAN once. That same SAN was up and running for over 5 years without a reboot and was just recently decommissioned. I also had this happen to me on some Cisco 26xx routers.
This would not happen regularly but I also always updated firmware on all equipment that I had used. I can say that I have had to reboot a Windows box with more frequency back then but not as much now.

The routers and servers now have some of the code written for the kernel or OS are written a little better. With the exception of Windows patches I do not have to reboot near as much. Most issues I deal with are related to poor security/programming by a developer.

blargoe

I'll go as far as to say the only time I ever reboot a server outside of a normal mainenance schedule is if the system isn't responding or has a hardware issue that requires the system to be down to remediate, unless getting to a clean state is the last resort to continue troubleshooting the root issue.

it_consultant

Windows won't let certain things happen without a reboot. Ever had to delete the "PendingFileRenameOperations" flag? Some of it is by design. You could mimic a reboot by restarting services, unloading and reloading DLLs etc - reboots, especially in a virtual guest, are easier and faster.

Zartanasaurus

I can assure you that 99.999% of the time, the root cause isn't "server hasn't been rebooted enough times".

it_consultant

"Did you reboot three times?" from "Sales guy vs web guy"

The Website Is Down

healthyboy

yeah,

like i said previously on my post, ofcourse a reboot is the last option but where i work sometimes servers or routers or switches just hang, meaning you cannot connect to them the only option is a reboot,

blargoe

it_consultant wrote: »

Windows won't let certain things happen without a reboot. Ever had to delete the "PendingFileRenameOperations" flag? Some of it is by design. You could mimic a reboot by restarting services, unloading and reloading DLLs etc - reboots, especially in a virtual guest, are easier and faster.

Well yeah, but usually something like that is caused by installing software or making a change that probably should be scheduled during a maintenance cycle... not always but usually.

Akaricloud

What causes them to "hang"? If it's OS or software then track it down and get it fixed. If it's hardware causing the problem then troubleshoot, find the cause and fix it. Constantly rebooting them isn't a valid course of action in my opinion.

There is no way I would ever let something like that fly with the servers that I manage.

SteveLord

More like "some desktop issues."

Once in a while, it is just easier to reboot as opposed to slaving over the issue. Unless of course...it happens again.

Tackle

I'm really suprised no one has quoted the IT crowd.

User issues - I have no problem asking them to reboot. Server issue - reboot as last option, usually after hours with someone on site watching it incase it requires user action to proceed.

Hello, IT. Have you tried turning it off and on again? Uh... okay, well, the button on the side, is it glowing? ...yeah, you need to turn it on... uh, the button turns it on... yeah, you do know how a button works don't you? No, not on clothes.

healthyboy

yeah, ofcourse reboot is my last restore but what do you do if you cannot access da server???? After i reboot i always check da logs to see what was up

bermovick

Some of the techs at work seem to think routers work like windows 95 boxes and actually suggest rebooting once it's been up a while without one (like 3-4 months). Very frustrating.

Akaricloud

healthyboy wrote: »

yeah, ofcourse reboot is my last restore but what do you do if you cannot access da server???? After i reboot i always check da logs to see what was up

That depends on what is in the logs. Are you actually doing anything to fix the cause of problems other than just looking at them? -If you're fixing the root of the problem then this should be a very, very infrequent event.

NetworkVeteran

like i said previously on my post, ofcourse a reboot is the last option but where i work sometimes... routers or switches just hang, meaning you cannot connect to them the only option is a reboot,

I've worked on routers and switches on a daily basis... it's exceedingly rare that business-class devices "just hang" for no discoverable reason. A good engineer will figure out why and solve it.. or if they lack the expertise and can't pick it up, call the vendor for help.. or if the vendor won't help, make a business case for replacement. If you're being paid and the network has a business need for existing, flaky equipment is costly in terms of lost productivity/opportunities and IT support costs.

blargoe

healthyboy wrote: »

yeah, ofcourse reboot is my last restore but what do you do if you cannot access da server???? After i reboot i always check da logs to see what was up

Of course u reboot da srvr if it locked up dogg!

blargoe

SteveLord wrote: »

More like "some desktop issues."

Once in a while, it is just easier to reboot as opposed to slaving over the issue. Unless of course...it happens again.

Desktops are a different animal in a way... if it's an issue that is isolated to one person's computer, you have to weigh whether their ability to perform their job is impacted by the issue and if the time required to troubleshoot at that particular point in time is acceptable, or if rebooting to get them back to doing their job is required. If it's a recurring issue, it's either find a window of time convenient for them to troubleshoot, rebuild their system to a clean state, or they just keep rebooting their computer. Often, it is up to the user and/or their manager.

the_Grinch

Actually, Citrix recommends a nightly reboot of servers running their software. Never could find out why, though I suspect it was because of hung sessions and the like. Way easier to just reboot them then to attempt to clear them, nothing worse then the "I can't get onto Citrix" at 8 am call. Then you track down the server, deny all logins, confirm they get onto a different server, and then wait for everyone to back off the thing so you can reboot. All while the IT manager at the company yells at you because she's down a server when they have 4 and are way below the threshold requiring 4. Also, on top of her not allowing you to reboot the servers nightly per the vendor Oh IT how I loath thy and love thy at the same time!