[Samba] winbindd crashing -- how to auto-heal?

[Robert Marcano]

On 09/01/2018 07:03 PM, Luca Olivetti via samba wrote:
> El 1/9/18 a les 03:03, Jamie Jackson via samba ha escrit:
>> Hi folks,
>>
>> Every couple days any one of my project's 10 or so RHEL VMs can't be 
>> logged
>> into with a domain account (active directory integration). Our admin goes
>> in with a local account and restarts winbind to fix it.
>>
>> I'd like to be more proactive and, say, create a cron job to detect a
>> problem and then restart winbind.
>>
>> Does anybody have a recipe for this? If not, do you have any 
>> suggestions as
>> to how I can begin to figure out how to detect the problem?
> 
> I run this cron job every minute. Winbind usually crashes once or twice 
> a day
> 
> #!/bin/bash
> getent group | grep -q 513 && exit 0
> echo "restarting winbind"
> systemctl restart winbind

If the failure is a full daemon crash and not some kind of internal 
failure, you can edit the service unit file (systemd) and add a restart mode

# systemctl edit winbind

   [Service]
   Restart=on-failure