Are you under the impression that servers don't get restarted to fix (or rather paper over) these kinds of issues? They most certainly do.
Most servers are running one or more applications as their primary purpose and if these malfunction then obviously it's generally the first port of call to just restart the app (when other troubleshooting avenues have been exhausted such as checking logs, talking to the vendor etc.) but if it's the OS itself that has an issue, yes, servers get restarted. Generally it'll be at a time that has the least or no impact to the consumers of the service. If any downtime is unpalatable then typically they will get clustered and the clusters will be split and restarted node by node. But servers, Windows and Linux, do sometimes require a restart.
[doublepost=1472580007][/doublepost]
No. Must've been before my time