Hi

 

i've been running unraid as a VM on my box "My711", and didn't have any major issues....i had few issues, but resolved them and it was because i added more disks than what the RAID card supports....beside that, never had any errors, even though i had several power outages, but the parity check ends successfully and no issues at all, and when i compare it to what i read to the problems people are having on the forum, i think i'm in great condition.

 

However, what are some tests and actions in order to verify the health of my system? i mean, i could be having problems and im not aware....

 

Also, how to test the full capability of my system in terms of read/write...etc.?

 

 

It is a good idea to monitor the smart reports on your drives. myMain (see link in my sig, shamefully promoted here) does a nice job of showing you the issues for all of your drives on one screen. It has a handy feature that is you ever seen an anomaly (like a reallocated sector), you can tell myMain to not report that anomaly again until the value gets larger. If you use this feature, your "smart view" should always show clean until something new occurs that requires your attention.

 

Another potentially good idea is to run the smart short and long tests. I have never done this myself, and feel that the monthly parity checks and checking the smart view from time to time allow me to keep tabs on the drives.

 

I would NOT recommend extensive cleaning and cable management that could result in cables getting shaken loose or disconnected. But do keep the dust bunnies at bay and clear any sort of air filters to make sure your case and drives have good cooling.

Make sure that the cooling fins on the CPU and other MB chips are clean.  Blow air through the fans and air inlets and exhaust vents to clear them out.  As someone mentioned a while back, computers can make excellent air filters!  You might want to do this in an area where the resulting dirt and grime would present a clean-up problem.

Temperature is not an issue with me, even though my case (Norco 4224) is in a wooden rack.

 

The max. temp i've seen on the drives are in 38-39, that is when i have a parity check or parity rebuild.

 

monitoring my case via IMPI, components are in the 32-34 range most of the time, except when i shutdown the A/C as I get cold lol.  When AC is off, it goes up to 38 degrees.

 

I haven't installed any addons on my unRAID since the first day, but I like MyMain, i might give it a try later.

 

 

and when i compare it to what i read to the problems people are having on the forum, i think i'm in great condition.

 

lol.  I know that comment is not directed at anyone but I feel like I fit the bill.  I've had several drives fail in the last year.  But... I love unraid.  Keep in mind that my server was bought directly from Lime Tech in 2007 and its only now after 7 years of continuous running that I'm having issues.  I had zero issues for many years.  The drives that failed are very old as well.  I think I have about 3 left from that age group.  I have one bad slot as well that failed a couple months ago.  All others are full.

 

The only time so far that I lost data was when I did something stupid one time and got impatient.

 

All that said.  I still love unraid.  I was considering a brand new setup bought fully from lime tech to act as a backup to this server till I saw the cost in Hard drives alone would kill me.  I guess when you add a drive or two at a time it really adds up over 7 years.

 

Current plan.  Wait till they have an official final ver 6 out.  Then buy a new server from them with about 3 very big drives.  Then I'll migrate my data over from the larger drives across the network.  As each drive gets emptied I'll move it over as well.  I doubt the last 3 orig small drives make it over.  We will see.

 

I'm also doing Crashplan now to backup offsite.  I don't run it on the server but rather my desktop.  A simple batch file called by the scheduler on bootup uses system to map my unraid shares so that Crashplan can access it.  Normal mapped drives don't work.  It works great.  I don't have unlimited bandwidth so it will take awhile before that is fully in place.

 

As far as maintenance.  Using unmenu for easy access to some of the plugins you can sched a monthly parity check to find issues with your system so you can find that failed drive and replace it before a second one goes.

 

Cal.

 

My newest addition for maintenance is keeping an md5sum file for each disk.

I update the missing files every few weeks. Do a full test once a month and update the files that have changed.

 

My next project is to keep a folder.par2 file per directory to recover from minor corruptions or bitrot.

 

Other then that a monthly parity check and some sort of scheduled full surface scan with badblocks in read mode or a SMART long test should be done.

 

I like to do the SMART long test as it puts an entry in the self test log and the hour of life it was run.

You can then track when problems appear/disappear.

 

@Calvin, if you start making those md5sum files now you will be ahead of the game when you start copying them over the network.

In addition, I've found some kernel tunings and rsyncd.conf tunings on the rsync server that can provide good performance at the command line/network rsync level.  Then there is always teracopy with windows which checksums the files before and after

 

I'm also doing Crashplan now to backup offsite.  I don't run it on the server but rather my desktop.  A simple batch file called by the scheduler on bootup uses system to map my unraid shares so that Crashplan can access it.  Normal mapped drives don't work.  It works great.  I don't have unlimited bandwidth so it will take awhile before that is fully in place.

 

 

 

Hi Cal, how do you do this? I faced the same issue with almost all the online storage similar to Crashplan, they can't backup mapped network drive.

 

Will be useful if you tell us the workaround

 

thanks

Here is the page I used to figure out how the batch file should be setup and run.

 

http://support.code42.com/CrashPlan/Latest/Backup/Backing_Up_A_Windows_Network_Drive

 

 

 

 

Here is my batch file with username and userpassword replaced but so that you can still see were they go.    Set it up with your ID and according to the notes in the link above.  This only works on boot.  You can't just run the batch and have it work.  It will fail.  It must run as "system" before you login.  The above links shows hows thats done.

 

You may not need the map delete command but I found it works perfectly everytime if I do it this way while I had some issues once in awhile if I did not delete them first before resetting them.  After you boot, it does take a minute or two for them to show up.

 

You end up with mapped drives that look like they are disconnected.  But you can still navigate them fine.  Depends on what your doing it will ask you stuff if you moving a file that it would not otherwise ask if you were using a standard mapped drive.  So... I don't use these drives myself.  I just use shortcuts to the UNC path instead of the mapped drive letters when I want to access the shares.  The mapped drives however are used seamlessly by crashplan as its running as a service itself.

 

 

 

echo %date% %time% : "%cd%\mapdrive.bat" >> C:\tools\bat\Crashplan\mount.log

net use * /delete /y

net use X: "\\tower\Avi - Keep" /USER:tower\username userpassword >> C:\tools\bat\Crashplan\mount.log

net use Y: "\\tower\Backups" /USER:tower\username userpassword >> C:\tools\bat\Crashplan\mount.log

net use Z: "\\tower\Share 3" /USER:tower\username userpassword >> C:\tools\bat\Crashplan\mount.log

 

 

 

Hope that helps.

Cal.

 

 

Edit. 

Ironically.  I used Crashplan at work to recover that batch file from my home pc backkup so that I could post it.  lol.

I could have taken it over to grab the file but this was easier.