How to monitor for failed HDDs

[Bruce123]

Hi,

 

I just wondered how people are monitoring for failed (or on predictive failure) HDDs on servers?

 

Are you able to do regular visual checks (looking for amber lights)?

 

Do you have a central system to monitor all servers? (E.g. Solarwinds)?

 

Do you use the hardware monitoring in VSphere/Centre?

 

Has anyone setup IBM MegaRAID software to monitor multiple IBM X series servers?

 

What about HP Proliant servers?

 

As regards to RAID setup, do you tend to have a hot spare on each server?

 

Kind Regards.

 

Bruce.

[speckytecky]

Dell Servers indicate straight away when one is going to fail. Nice little blue window turns orange with a message. Taking the front bezel off shows the guilty drive again with an amber light instead of the usual green. Then there's the excellent Dell Open Manage that shows more detail. We have 4 hour response and using that have submitted a Dell Warranty case at 0830 hours in the morning with the disk supplied, fitted and rebuilt into the server by midday. Fantastic service.

[Joanne]

Dell Servers indicate straight away when one is going to fail. Nice little blue window turns orange with a message. Taking the front bezel off shows the guilty drive again with an amber light instead of the usual green. Then there's the excellent Dell Open Manage that shows more detail. We have 4 hour response and using that have submitted a Dell Warranty case at 0830 hours in the morning with the disk supplied, fitted and rebuilt into the server by midday. Fantastic service.

 

Yeah mine is the same... don't have that 4 hour service though! I get one shipped out next day. I check physically once or twice a week. I always go in the room to make sure people haven't left a load of paper on the server - it's in with the photocopiers. I also use the open manage software.

 

Other ways to check are event logs. You will get a lot of NTFS errors. Normally points to a potential failure.

[SHimmer45]

Dell HP IBM will all have monitoring agents which will report faults if you login and look at them or set them to send an SNMP broadcast or install the management console (in Dells case its OME/Open manage essentials) which gathers all the reports from the agents and then takes action as required.

 

Hot spares are useful but you imo need the alerts to say its got a problem as you dont want a situation were a disk fails and then another go while its trying to rebuild on to hot spare.

i keep a selection of cold spares for when its required

[jamesfed]

PRTG watches the HDDs on our Dell PowerEdge servers through the iDRAC interface -

 

 

Easy to setup and works a treat; if anything goes wrong we get email alerts and a text message

[DavR]

+1 for iDrac / chassis controller monitoring.

 

All of my Dell servers plus my storage array are all set to email me on warnings and failures. Touch wood I've not had a bad drive on my current setup yet though, so I only know this works in theory so far!

[DavR]

Also, in a previous life, I had CentraStage, which could do Event Log monitoring.

 

If the correct vendor tools (OpenManage Server Admin etc) are installed, when you get a bad drive, the tools report this as error or warning in Event log. CentraStage then sees the error and creates an alert.

 

If was with great trepidation that I first turned this on across something like 120 servers across 60 customers - that generated a whole bunch of drive replacement jobs! Got me out of the office for a few weeks anyway.

[ITGURU]

Full HP Proliant suite of servers here - all monitored through FREE HP Insight Manager - as soon as a NIC loses connection, PSU fails, HDD fails etc etc I get an email, and shows up in the web interface.

Just enabled the onboard NIC as not used to show example. but similar notifications for other component failures.

 

 

No hot spares, but always have a couple of drives in stock readily available to replace.