ManishGupta8
-
Posts
60 -
Joined
-
Last visited
Content Type
Profiles
Forums
Downloads
Store
Gallery
Bug Reports
Documentation
Landing
Posts posted by ManishGupta8
-
-
I did enabled the logging and set cache yes...It says mover is running but i dont see any file in my main data disk/.
-
I am on Unraid 6.6.6 and I ahve 2.5 TB on my cache drive. I clicked on "Move now" button under "Main" and it is still not moving data into my main data drive.
Do i need to do anything else as well ?
-
I recently created new array and performed preclear. after adding the disks I am getting orange triangle on data drive and not sure what's wrong there ?
Can someone please help and explain ?
-
Thanks Johnie. You are awesome.
-
19 minutes ago, johnnie.black said:
64MB/s average is still rather slow, but you can try using that disk in the array together with the previous fastest one, if writes speeds are now around 60MB/s instead of 25MB/s you know the disks are the problem.
So far i have done preclear only 2 times on Parity and 1 time on Data Disk (Disk 3). Before I add them to array do you suggest to do preclear again 22-3 times ?
-
On 12/15/2018 at 7:40 AM, ManishGupta8 said:
Interesting...I Started Preclear on both 8 TB Disks that i got from Costco at same time...Currently both are running (in N54L with 8 GB RAM) but i can see huge difference in speed. Preclear is not yet completed but see the current status....
Hi Johnie and Turyl,
Here are the Stats when i ran preclear on new 8 TB Disk. Still Slow as compare to 1 st preclear result but much better than previous 8 TB Hard Disk.
Should I run preclear again on both Disk 1 and Disk 3 together again 2nd time ?
-
Just now, johnnie.black said:
remove the slow disk from server, install a different one, see if write speeds are normal now.
Thank you. Will do so and share the results.
Once again THANK YOU VERY MUCH.
-
Just now, johnnie.black said:
I am confused with that statement. What do you mean ?
-
2 minutes ago, johnnie.black said:
Preclear has been known to show wrong stats and/or clear disks at different speeds when clearing multiple disks, though the results appear to confirm your issue nothing like replacing the slower disk and test to confirm.
Thanks. So are you suggesting to cancel currently running preclear and do the same again but this time only on the disk which is running slow as per preclear ?
-
6 hours ago, johnnie.black said:
It's possible, if the preclear stats are correct.
Hi Johnie,
What do you mean by if preclear stats are correct ? Is there any possibility of same as well ? I attached preclear results so i believe i am reading it right.
Please help.
-
On 12/14/2018 at 8:39 AM, johnnie.black said:
No, we are waiting for you to test with different disks, like you said.
Hi Johnie and Turyl,
As you suggested I have done some test and I was wondering if you can share your inputs on same. Below is a brief summary of what i have done so far.
- Purchased two 8 TB Seagate Hard Drives from Costco At Same time.
- Both are connected to Proliant N54L with 8 GB RAM exactly in same way via SATA Connector (both connected to Motherboard via same SAS connector)
- There are no other Hard Disks connected to Proliant machine as of now except these 2.
- Unraid ver 6.6.6 was installed alongwith Community applications and PreClear Disk (from gfjardim's plugin Repository)
- I started preclear on both hard disk almost at same instant.
- Disk 1 Took in total under 47 hrs to finish preclear. Out of these 47 Hrs it took rought 15 hrs for Step 2 i.e. "Zeroing the Disk" at average speed of 144MB/sec. Screenshot attached below.
- Disk 2 has already taken 66+ hrs and still preclear is not finished. Out of 66+ hrs it has consumed 51 hrs in Step 2 and still only 58% completed. Step 2 is running at average speed of 25MB/sec.Screenshot attached below.
I am feeling as if I have just got one of the bad hard disk (Disk 2) which is causing this issue but other hard disk (Disk 1 ) is correct.
Can you guys please share your views on same ?
-
22 hours ago, johnnie.black said:
No, we are waiting for you to test with different disks, like you said.
Interesting...I Started Preclear on both 8 TB Disks that i got from Costco at same time...Currently both are running (in N54L with 8 GB RAM) but i can see huge difference in speed. Preclear is not yet completed but see the current status....
-
6 hours ago, johnnie.black said:
Precelar only writes zeros, and that can be handled different by the firmware, i.e., not hit the SMR wall, you should just test with the other disks, it would confirm if the issues are the disks nor not.
Currently the OP is using a single data disk plus parity, so Unraid works as raid1, no parity speed penalty.
So Are we concluding that its definitely Hard Disk fault ??
-
5 hours ago, johnnie.black said:
Yes, they have a multi-tier cache, usually comprised of RAM -> Flash -> PMR and only then SMR, what's interesting is that the old Seagate Archive drives were very good at detecting sequential writes, and as long as you were writing large files they kept up with good speeds, for example, this is a server with only 8TB Seagate Archive drives:
Total transfer size was about 150GB of large files, and speed remained constant, but not all SMR drives can do this, for example I have a few WD 2.5" 1TB SMR drives, and even for sequential transfers with large files they are only fast for the first 50GB or so and then write speed decreases to around 25MB/s, and possibly the newer SMR Seagte Barracuda drives behave similarly, which would still be kind of strange since the Archive drives work so well, but like I mentioned there are other reports of these models being slow.
Somehow i feel there has to some other reason. I started preclear as someone suggested in Seagate * TB thread and just wanted to check the speed. Currently it is 3% done but it is average speed is showing as 180MBPS. If preclear can run at this speed then why not my standard writes from windows to Unraid over LAN CAT6 cable ? My file sizes are also not just 1 GB rather generally they are of size 2-3 GB in general.
-
On 4/8/2018 at 3:20 PM, HellDiverUK said:
The Seagate external will likely be a Barracuda, which is a SMR unit. They work fine in unRAID.
The general recommendation is the WD externals which usually contain a WD Red (or the white label equivlent).
I have tried many combinations and cant get more than 25 MB/S out of Seagate Baracuda.
Refer below thread and if by any chance you can suggest anything then it could be game changer for me. Else all my money spent on these 8 TB drive is waste now.
-
28 minutes ago, johnnie.black said:
New config (Tools -> New config) makes Unraid "forget" current assignments, after doing it you'll need to reassign the disks, but nothing will be deleted on the disks assigned as data drives, if they have data and were formatted by Unraid or UD it will still be there, starting the array after a new config will re-sync parity, i.e. create valid parity, you just need to be careful and don't assign a disk with data to the parity slot.
Johnie, Do you think it could be because of underline 1 SAS connector which is gettign split into 2 SATA Connectors (1 for Parity and 1 for Data Drive) ??
-
1 hour ago, whipdancer said:
I'm not asking what a data drive is. I'm asking what kind of drive are you specifically using for your data drive?
You cross posted this question all over the place - did you read the Seagate thread you posted on? It has an answer for you.Hi Whipdancer,
I was just wondering if you know the reason for slow speed. You meantioned that other thread has an answer but I don't see anything there . Seems everyone is reporting all good with "Model Family: Seagate Barracuda Compute and Device Model: ST8000DM004-2CX188" hard drives.
-
1 minute ago, johnnie.black said:
New config (Tools -> New config) makes Unraid "forget" current assignments, after doing it you'll need to reassign the disks, but nothing will be deleted on the disks assigned as data drives, if they have data and were formatted by Unraid or UD it will still be there, starting the array after a new config will re-sync parity, i.e. create valid parity, you just need to be careful and don't assign a disk with data to the parity slot.
Thank Johnie. I will do that. Good to know. I know we drifted from original post and it did not get solved but everything i learn here is good for me
-
1 minute ago, johnnie.black said:
That's a good plan.
Yes, but only if there's no parity assigned, or by doing a new config and re-syncing parity.
Ok let me try to do it and will update others the o/p of same.
Is there any link/video available where i can learn what do we mean by new config and resyncing parity ?
-
2 minutes ago, johnnie.black said:
It fits with the problem you're seeing, i.e., the speeds starting normally and then slowing down once the PMR cache is exhausted, and like mentioned other users suffered from low performance with those same disks, still it's only a guess, to be sure you'd need to test with other non SMR disks, or at least SMR disks that are know to perform well, like the Archive drives.
¿Que? It was already explained by trurl why cross-posting is bad, and you already posted the diagnostics...
Thanks. I am not sure why that old comment also came in when i clicked on "Quote" button. But yes truyl already explained that part. I am thinking to migrate my old server Bootable flash and all disks to new server to ensure that issue is somewhere in drive. In both machines all is same except the hardware and hard drive. So if old hard drives (from old proliant N54L) works at great speed in new proliant ML10 then it clearly mean it is an issue with hard drive . What do you think ? Please share your views ?
Also is there any way i can add a brand new data disk with data on it to array without formatting. (Assuming it already is in format "reiserfs")
Actually thinking to do reverse as well then. Connect 8 TB drives to old server using unassigned and then copy data, so that my old array is not down for long. Once copied then move this new 8 TB drive to new server , add it to array (Without formatting it) , but rebuild parity. Not sure if this is even possible.
-
1 minute ago, johnnie.black said:
If that's the problem only changing the disks would help.
Thanks for trying to help. Dont get upset but as i dont know so will rely on what you said. So is this confirmed based on some diagnostic logs i shared or are you guessing that could be the reason based on your pas experience. ?
30 minutes ago, johnnie.black said:Please don't crosspost, and post your diagnostics, it might show something and if nothing else it will show more info about the hardware.
Hi Johnie. I did not understand what do you mean by Crosspost. So please explain so that i can avoid it.
Also how can i get disgnostics please ?
-
1 minute ago, johnnie.black said:
Looks like you're hitting the SMR wall.
Can you please elaborate what to change and how to change. I am not expert and very basic user of Unraid. Just use it for my media storage nothing more.
-
3 minutes ago, johnnie.black said:
Nothing jumps out, but there have been other reports that those newer SMR Barracuda drives perform worse than the older Archive drives, I never used one myself, but have some of the Archive models and those perform great.
Iis the low speed immediate after a copy start or does is start faster and then slow down after a few GBs?
Slow speed is not immediate...It starts around 150MB/S and then start reducing and get stabilize at 20MB/s
-
1 minute ago, trurl said:
Posting the same thing in multiple places is known as "cross-posting". It has be frowned upon on message boards since long before the World Wide Web.
I deleted another new thread you started, and I split the reply another user made in the other thread into this thread.
When you post the same question in multiple places, where are we supposed to respond? What if someone goes to a lot of time and effort to research your problem and writes a nice, long, detailed response. And then another person, not aware of that thread, goes to a lot of time and effort to research your problem and writes a nice, long, detailed response on the other thread, totally unaware that they are wasting their time because they never saw the other thread.
Cross-posting is bad because it makes it impossible for the responses to be coordinated.
I apologize for same and what you explained makes sense.
Manually run mover script for Cache Drive
in General Support
Posted
My unraid keeps on saying
"Please wait... collecting diagnostic information (this may take several minutes)" ... It has been over 15 minutes now. Does it take so long ?