# Should answers encouraging to “restart” be flagged?

Lately, reviewing the queues, I'm seeing a lot of answers encouraging the OPs to simply restart the service that isn't working or even their computer/server to see if that solves the issue.

It's quite clear that these are not good answers, so I usually flag them as Not an answer or Very low quality.

The problem comes when some of these answers are upvoted (some of them even with more than one positive vote), or even accepted (usually the OP replying themselves arguing that they finally managed to solve it by restarting the mentioned service/server/whatever).

• Should these answers be flagged, it doesn't matter how many upvotes they have?
• In the case where the own question author replies themselves with the restarting worked for me, should we vote to close the question as well?
• Statement: I'm also flagging answers containing restart advice as "Not an answer". And sometimes when answer contains "restart" witch hunt begins meta.superuser.com/questions/10961/… – g2mk Jan 5 '16 at 13:14
• Unfortunately sometimes those low quality answers actually solve the issue, but they should be a comment if they are one or two liners imho. – Moab Jan 9 '16 at 16:48
• why is messenger guilty that half the issues are resolved by simple restart </meducks> especially on proprietary operating systems where you have little ability to debug/fix the issue or even report it. – akostadinov Jan 15 '16 at 19:57
• I think rebooting/restarting is an acceptable method for some problems. The longer programs run, the more possible states they could enter leading to unknown behavior. In the absence of debugging programs and the program's source, it's impossible to troubleshoot. This is especially true regarding certain combinations of drivers and hardware that are difficult to reproduce otherwise. – nijave Jan 17 '16 at 22:10

Good question. However, I'm going to play devil's advocate and explain why I think answers like this are fine.

Let's generalize the concept to "simple but effective remedies". This is a category of answer that's extremely easy, well-known, and solves a broad variety of problems, either temporarily or permanently.

My opinion is that these simple but effective remedies should not be posted as an answer, unless the answerer is pretty-darn-sure that it'll actually resolve the problem, due to personal experience or professional advice (e.g. if a KB article says to do it, then it's probably effective. Probably.)

So consider two types of answerer:

1. The clueless user trying to milk easy rep by spamming every question with "Try rebooting, that might fix it".

2. The well-reasoned researcher, who googles around (or has personally hit the same problem) and finds that a reboot or other simple common remedy is the most likely fix. They also provide a link to any evidence they have to third parties that have said that this remedy is effective.

In the case of answerer #2, I'm fine with them posting such an answer, and we shouldn't delete it. In the case of answerer #1, hopefully they get enough low quality flags that the answerer gets suspended or answer-banned after some time.

I feel that the quality control systems already in place provide sufficient filtering of bad answers that simply prompt the querant to reboot without providing any details (especially for questions where doing this is not actually effective).

I also think there are also probably sufficient systems in place to prevent good answers (with reasoning, links, etc.) that prompt the user to do this with good reason from being deleted.

Therefore I don't think any changes are needed to account for this type of answer. We're already doing what should be done in most cases. If you find a specific instance where this was handled badly, e.g. a bad answer saying "Try rebooting, see if that helps" in response to a "How do I format my cells in Excel?" question, flag it. If you find an answer where it's downvoted or deleted and the "Just reboot and it'll fix it" answer is actually correct, flag it to be undeleted and/or upvote it.

• I agree in principle provided there is more to the "restart" advice the. Just saying to restart, and while I realize some problems are really that simple but there has to be something that can be said. Anyone can say "restart" a process – Ramhound Jan 5 '16 at 9:24
• Let's assume a #1 answer gets accepted, in this case we would be able to close it as "went away" if we would have this close reason as available on SO and Askubuntu. (#2 would open this option too, but would not have to be enforced with a high quality answer) – bummi Jan 16 '16 at 16:28
• I think it is also important whether the reboot is going to solve the problem permanently. If a reboot will make the problem go away temporarily, but it is bound to show up again if the system is left running, then I wouldn't consider an answer suggesting to reboot to be a valid answer. – kasperd Jan 17 '16 at 21:51

I can't imagine a situation where an answer suggesting to reboot might be useful to a reasonable user. In fact, the first thing reasonable people do before asking is making sure the issue is reproducible, and that should include rebooting. Unfortunately, some people rush to ask, so this catchphrase never gets old.

I'd say if the answer suggests to reboot (and nothing more) it should be deleted or converted to a comment, even if the answer has helped the OP. Actually, if the OP accepts such an answer, it's a strong indication that the question itself should be closed as "Not reproducible".

Of course, this does not apply to answers which suggest a reboot as one of the steps, to make sure the OP doesn't miss it and does reboot at the right moment.

• What about "Make sure that it's plugged in" as an answer? :-) – fixer1234 Jan 6 '16 at 8:18

What about the possibility that the computer is behaving erratically, and a restart is the only way to clear the memory of the all the various 'data junk' that accumulates over time?

The benefits of shutting down your computer according to SuperUser contributor David Zaslavsky:

From a software perspective, an operating system and the programs you run on it tend to accumulate all sorts of cruft over extended periods of use – temporary files, disk caches, page files, open file descriptors, pipes, sockets, zombie processes, memory leaks, etc. etc. etc. All that stuff can slow down the computer, but it all goes away when you shut down or restart the system. So shutting down your computer every once in a while – and I do mean actually shutting down, not just hibernating or putting it to sleep – can give it a “fresh start” of sorts and make it seem nice and zippy again. However, different computers and OS’s are not all equally affected by this phenomenon. Generally, a computer with a lot of RAM can go for much longer than a computer with only a little RAM. A server, on which you just start up a few programs and then let them work, will be fine for much longer than a desktop computer, where you’re constantly opening and closing different programs and doing different things with them. Plus, server operating systems are optimized for long-term use. It’s also been said that Linux and Mac OS tend to run for longer than Windows systems, although in my experience that mostly depends on what programs you use on them, and not so much on any differences between the kernels of the operating systems themselves.

• It isn't so much whether it works. Restarting is often a good solution for a broad range of issues. In fact, that's why it's often considered a "joke" response. There are exceptions (see allquixotic's answer), but most questions for which a restart is the solution are probably not a good fit for the site. See Dmitry's answer. They tend to be not reproducible, or too broad in terms of possible underlying problems. If people don't check basic stuff like that before asking here, we'd have thousands of questions all with the same answer, and people wasting time trying to diagnose hiccups. – fixer1234 Jan 16 '16 at 1:52