I am a new owner of a ready nas 100-4 so I'm not really expert on this platform as so far I always owned and worked with different hardware than this.
I installed my disks (3x3TB WD-RED) and started the initial sync of the data volume. Despite beeing unhappy with the performances (around 60 hours to complete the process) the sync went fine without any error.
NOTE: at this point I didn't copy ANY single file on the nas due to the ridiculus performance of I/O operations during the resync
While I was away this weekend a power outage happened in my house, so I managed to power on again my nas only this morning. Well guess what? The resync process started all over again without any single log explaining WHY is this happening.
No I/O was happening during the outage, ALL the sharing services were STOPPED to let the sync go as fast as possible so the machine was doing absolutely nothing (expecially no I/O operation on the data volume).
So my question is: is this always going to happen? I'm ok to resyncs, but during the 60 hours of that job, it's totally impossible to use the nas as the CPU stays at a fixed 100% usage by sys process and any read or write operation is just as painful as using floppy disks.
P.S. Please don't tell me "get an ups" as this is not the point. I really need to know if this is something can happen again as I can't afford useless resyncs. It must happen ONLY when I get my raid in a degraded state, not in random cases. If after every power failure or other situation I didn't discovered yet the machine will make a resync, I have to return it immediately as this can't fit my needs.
Also, this are the log entries (sooo verbose....) Mon Jan 25 2016 12:19:14 - Volume: Resyncing started for Volume data. Mon Jan 25 2016 12:18:52 - System: ReadyNASOS background service started. Mon Jan 25 2016 12:18:48 - System: ReadyNASOS service or process was restarted. Fri Jan 22 2016 21:31:22 - Volume: Volume data is resynced.
The RAID is at a lower level than the filesystem and the disks need to be synced sector by sector to setup the RAID. Please note that the initial sync is a special case where the volume is redudant throughout.
The 100 series products are cost effective. We have higher performance options available.
Which firmware are you running?
Which services and apps are you running?
The system is designed to handle unsafe shutdowns. After some unsafe shutdowns there may be some concerns about the state of the array, so as a precaution in such cases a resync is triggered to make sure the disks are in sync. This would be advisable regardless of what system you are using. We want to minimise the risk that a single disk failure would result in the array being lost. The redundancy of RAID-5 is only effective if the disks are properly synced. If for any reason they are not then a single disk failure may cause problems.
Of course important data should be stored on multiple devices. If the only copy of your data is on the NAS then that data is not backed up. Sadly some users have misconceptions as to what benefits RAID provides. See Preventing Catastrophic Data Loss
The suggestion to get a UPS is a good one. If you get a UPS and connect it to the NAS via USB (in addition to plugging the power cable from the NAS into the UPS, of course) then the unit can be shutdown gracefully by that in the event of a power failure.
The system was shutdown improperly, and it does detect that. You say that no writes were pending (and I am not challenging that), but the system doesn't have any way to know that for certain. It is prudent for the system to perform file system integrity checks and resync the array in this situation. I suspect other NAS from other vendors would behave similarly.
Of course you already do know what to do (that is, buy a ups).
I understand the raid happens on a lower level than a FS, and for any I/O operation the sync keeps happening in the background. The thing is that there wasn't any I/O operation happening since the sync finished as I was not at home.
The machne was completely IDLE.
To make the sync performing better, I stopped every service except SSH and HTTP to continue to manage the device so at the moment, while it's resyncing again, the only running services are SSH and HTTP.
My firmware version is 6.4.1 and should be the last one.
Is there any way to understand why this resync happened? Is there any log I can check? Thanks
I agree regarding the file integrity check. After an unexpected shut down you might have corruption on files. The fileystem itself can actually have issues, needs to apply the journals and so on.
What I don't undertand is the whole resync thing: It doesn't make sense to me that the machine, which sees the raid in healty status, start resyncing it. Also, I have no clue about what's actually resyncing and from WHERE! If you break an HD and you replace it, the resync process takes data from the others to that new hard disk. But in this case I have no idea what's going on and why. The filesystem is still intact though, so it's not recreating the array from scrach but just syncing.. which I can't properly understand what does it means at this point!
Every time you write to the volume, you not only write the data, but you also need to update the corresponding RAID parity blocks. Both of these need to be done as an atomic operation.
If there is a possibility that writes were lost, then of course it is also possible that the parity blocks are out of sync with the data (e.g, the data write completed, but the system went down before the new parity block was written). If that happens, then later on (when a disk is replaced), data corruption will occur, since the parity blocks are used to recreate the data on the new disk.
You deal with this possibility by checking the parity blocks and repairing them. With mdadm, this is generally done with checkarray (and after the file system check), but checkarray is indistinguishable from resync to the kernel.
