tomba
-
Posts
13 -
Joined
-
Last visited
Posts posted by tomba
-
-
Tomba, did it say "scanning" or "waiting to scan"? Because normally, if there is disk activity, Scanner will delay the scan, as to not interfere with normal file operations.
That, or, have you messed with that setting? Or, it was possibly low enough activity where it didn't trigger this delay in Scanner.
It was definitively scanning, saw this both in the dashboard and resource monitor (100MB/s scanning of DR1)
Which setting do you mean exactly.
Anyway, an update on the time everything took.
- Disabling duplication on the biggest directory (2972GB) to allow the 2 TB to be removed: 2 hours
- Removing the 2TB disk (which had about 1TB of data left): 7 hours
- Re-enabling duplication on the directory to duplicate it to the 4TB: 7,5 hours
Specs of the server:
Dual Core Pentium Core2Duo E2220 (2,4GHz)
2GB of RAM
128GB Crucial M4 SSD (bootdisk)
1x Toshiba 3TB 7200 rpm
2x Hitachi 4TB 5900 rpm
The 2 TB disk was a 5400 rpm Samsung disk
-
I'm just as guilty of the "not duplicating everything". I'd like to, but ....
Also, you could use the "File Placement Limiter" balancer to force migration off the disk in question. That way, you don't have to sit waiting for the drive to remove. You can clear the data off and then quickly remove the disk.
LOL, saw this a little bit too late
The removing of the 2TB (containing about 1TB of data) took about 7 hours. I then placed the 4TB drive and since then Drivepool's been working for about 13 hours and is currently at 94,0% (Total disks in DrivePool; 3, 1 3TB and 2 4TB's)
One thing I did notice is that StableBit Scanner started scanning concurrently with the duplication (because it's new of course). I would have expected it to wait with the surface scan until DrivePool is ready.
-
Well, if everything was duplicated... you could have just pulled the drive, put in the new one, and then wait for it to duplicate to the new drive. That would have been the fastest way.
Well not exactly everyting was duplicated but hey
Thanks for your input!I have decided to go the 'normal' way, removing and adding. I will post the times needed here when I am finished!
-
Yes, as Drashna says, DrivePool should accept a cloned disk, just do NOT let it see both disks at the same time until you've removed the poolpart from the original disk.
I'm curious, why not just start it as an overnight job before you go to sleep? That it will take hours shouldn't matter then. At 50MB/sec it would've taken under 12 hours, or less than the time between when you first posted and when Drashna replied.
Hmm. Do you happen to recall exactly how many hours it took to remove your 2TB disk last time? Wondering if we should establish some benchmarks for removing disks (e.g. to help answer questions like "hey, it's been fourteen hours and my 2TB disk hasn't finished removing yet, is something wrong?").
Allright, glad that's settled. The reason to do it this way is mainly the sound (the server is near my bedroom) and costs (electrical). Also; the server is not very usable due to high diskload in the mean time
The removal of the 2TB disk took several hours, about 10 if I recall correctly, but it was filled to the brim.
-
Yes. As long as the drive provides the information via SMART. And a lot of the modern SSD include a "health" attribute of some sort which we monitor.
For instance, my OCZ Vertex 4 has some of the standard "wear" indicators (such as "reallocated sector count", and "Read Error Rate"). But it has a couple of specific attributes. Namely "Available Reserved Space", which measures the "secondary" data cells used when the primary ones wear (a good indicator of how much life may be left), as well as a "Media Wearout Indicator", whose name should be very intuitive.
Thank you for the elaborate answer
-
As long as both disks aren't in the pool at the same time, there should be no issue. But why not add the 4TB to the pool, and then remove the 2TB?
Namely, I ask, because that 2TB is likely formatted as "MBR", while that 4TB will most likely need to be formatted as GPT. (advanced formatting may not work when you clone the disks, which means you'd be limited to ~2TBs of space)
The reason to clone is, is to not have to wait for the pool to empty the 2TB. Did that a couple of months ago and it took hours upon hours. Trying to save that time.
But, if I understand you correctly it will work? (Cloning the data to the new disk will give me 2TB more space on the pool without having to wait for the data to copy out and back in?)
-
If I want to replace an existing 2TB drive by a new 4TB, will Drivepool be able to handle a cloned disk? (as in I ghost the existing disk to the new one)
-
Yes, you can definitely configure Scanner to send you email notifications.
Ok, maybe my question was a little bit ambiguous. I know I can get email notifications (already receiving those), my question is more: is Scanner able to detect a failing SSD just like HDDs, as in is it able to correctly interpret the Smart data provided by an SSD?
-
Yes, you can do this. Be mindful of the quality of the SSD you plan to use - most(?) have greatly improved, but the lifespan of an SSD is still ultimately proportional to the number of write/erase cycles, and feeder disk activity is basically a daily series of write/erase events.
And since the SSD will also have your OS on it, if your SSD's manufacturer has software that allows you to monitor the projected remaining life of the disk (e.g. Intel's Toolbox and Samsung's Magician software), I'd strongly recommend using it.
Thanks for your response, that's very good news
I have a Crucial laying around so no toolbox. -
Depends on what you mean by "monitors".
Scanner definitely grabs the SMART data from SSDs, as well as does the periodic file scans.
Will I get a warning when it goes awry? My current bootdisk is generating Smart Warnings which is leading me to think about exchanging it for an SSD.
-
As I am thinking about using an SSD feeder disk for Drivepool, I was wondering if StableBit Scanner also correctly monitors SSDs. Am I correct in asuming it does?
-
As the Boot drive of my WHS2011 server is slowly dying, I want to replace it by an SSD to gain some speed. Is it possible to use the remaining part of the SSD (all the space excluding the C:) as a feeder disk? (e.g. just the D: partition)
SSD monitoring
in General
Posted
I know how you feel
Personally I believe in credit where it is due...
The new disk I put in has 100% remaining of it's life, we'll see how it goes when (if?) I start using it as a feederdisk for DrivePool as well