D-Link Forums

The Graveyard - Products No Longer Supported => D-Link Storage => DNS-345 => Topic started by: psdont on April 16, 2013, 05:30:51 AM

Title: Serious logic flaw in firmware
Post by: psdont on April 16, 2013, 05:30:51 AM
I installed the DNS-345 a couple of months ago after being please with earlier DNS-323 products.  Before adding the drives, I updated the firmware to the current version 1.0.1.  Volume_1 is setup with two WD 4GB drives in Raid 1 configuration and Volume_2 is setup with two Seagate 3TB drives also in a Raid 1 configuration.  My problems began about a week ago with the first drive in the Volume_1 group was reported as failed.  Presented with the option to rebuild, I chose that and all appeared to be fine after the volume was rebuilt.  No reboot was performed.  First indication of a problem came that night when the daily email with the stats of the volumes listed both volumes but the name for the second volume appeared as "  ".  Although it was still possible to access the drive, I was unaware of the potential problem to follow.  After a couple of days, I began to experience network issues - after some diagnosis, it appears, the Volume_2 group was causing timeouts when it was attempted to be accessed.  That night, I shut down the unit to take a further look at it in the morning.  The next morning, only the Volume_1 group was recognized.  The two drives for Volume 2 were only recognized as "new" drives and the only option was to reimage.

Not being a programmer, I can only assume that the problem is with the logic for the rebuild of the initial volume - it wiped the table for the data on the drives in the unit.  Needless to say, the reimaging of Volume_2 resulted in the loss of data - approximately 20% of the drive was in use at that time.  Guess I will leave the Volume_2 group spinning in the unit and not put any data on it until this problem is addressed.
Title: Re: Serious logic flaw in firmware
Post by: JavaLawyer on April 17, 2013, 07:43:00 AM
Is there anything else you recall doing during the setup procedure that you may not have included in your post? I will forward this thread to my contact at D-Link after you've posted back with additional information.
Title: Re: Serious logic flaw in firmware
Post by: psdont on April 17, 2013, 01:12:17 PM
No I think that pretty much covers what I did to get access to the second volume group (Volume_2).  In short
1) Rebuild volume_1 following the "failed" message
2) Continued to operate the DNS-345 per normal (did not reboot after the rebuild)
3) Email of stats following rebuild listed the two volume groups (Volume_1 and Volume_2) but the name for the second volume was "  "
4) After about 36 hours, began to experience network slowdown / outages (isolated to the network segment containing the DNS-345)
5) Shut down the DNS-345 that night as the network issues appeared to be from timeouts when attempting to access the second volume
6) Reboot the DNS-345 the following morning
7) Only Volume_1 listed - the exiting 3rd and 4th drives (Volume_2) seen as new drives to the DNS-345
8) Performed setup of Volume_2 group (note, Volume_2 was approximately 20% utilized before the problem)

Oh - there was one thing more, attempted to post to the D-Link support board but after pressing the submit button received a message that the URL was not found - that's when I posted over here.

Thanks for forwarding.
Title: Re: Serious logic flaw in firmware
Post by: JavaLawyer on April 18, 2013, 08:17:32 AM
As promised, I forwarded this issue to D-Link.

Can you please perform a SMART scan for the HDDs in the RAID-1 that failed to re-validate their health?
Title: Re: Serious logic flaw in firmware
Post by: psdont on April 19, 2013, 03:44:32 AM
Here are the S.M.A.R.T. scans of the drives that had been configured as Volume_2 (Raid 1) and were not recognized after the rebuild on Volume_1.  I indicated above they were Seagates - those are in some of the other DNS's - these ones are Hitachi

A SMART Test Was Performed On The Following Hard Drive At 22:2:12 On 18-April-2013.

Device Model:  Hitachi HDS5C3030ALA630  
Serial Number:  MJ1311YNG1Z43A  
Size: 3,000G

The Result Of The Test Is: Pass

A SMART Test Was Performed On The Following Hard Drive At 22:2:15 On 18-April-2013.

Device Model:  Hitachi HDS5C3030ALA630  
Serial Number:  MJ1311YNG222JA  
Size: 3,000G

The Result Of The Test Is: Pass

Sincerely,
Your RMedServe
Title: Re: Serious logic flaw in firmware
Post by: JavaLawyer on April 19, 2013, 08:25:23 AM
Thanks. . . I'll pass this along as well. 
Title: Re: Serious logic flaw in firmware
Post by: JavaLawyer on April 24, 2013, 11:30:28 AM
D-Link had three questions for you:

1. HDD model name/number
2. Which HDD (slot position) was used for Volume_1, and Volume_2?
3. have you ever done hot-swap on these drives?  

Also sent you a PM.

Thanks!
Title: Re: Serious logic flaw in firmware
Post by: psdont on April 24, 2013, 01:11:17 PM
1) Two drives are Western Digital WD4000FYYZ-01UL1 and two drives are Hitachi HDS5C3030ALA630.

2) The Western Digital drives (Volume_1) are in slots 1 and 2 and the Hitachi drives (Volume_2) are in slots 3 and 4 (top to bottom)

3) No, the drives have not been hot-swapped.  They have not been touched since being inserted into the unit.
Title: Re: Serious logic flaw in firmware
Post by: JavaLawyer on April 25, 2013, 04:45:52 AM
Thanks....  ;)