November 8, 2017 at 3:51 am
Hum I see it, no basic the disk are very simple environment it is a Raid 1, it is nothing complicate, is a medium HP server.
So will be difficult to talk with the Vendor.
Thank you
November 8, 2017 at 3:56 am
Hi All,
That is the Point each Server is a Physical one with the same OS , Hardware and SW configuration standards.
So I already Check and looks for hardware evidences what could be broke or with issues and was nothing record in the Event Viewer.
So I no idea what could be the Problem since there is nothing I could say I got a Memory issue or a Disk or a Network.
That is what is frustrate me.
Thank you
November 8, 2017 at 7:22 am
d_gomes2 - Wednesday, November 8, 2017 3:56 AMHi All,
That is the Point each Server is a Physical one with the same OS , Hardware and SW configuration standards.
So I already Check and looks for hardware evidences what could be broke or with issues and was nothing record in the Event Viewer.
So I no idea what could be the Problem since there is nothing I could say I got a Memory issue or a Disk or a Network.
That is what is frustrate me.
Thank you
They may not have shared resources but from your description they have the same drivers, IO components, NICs, etc. Which means they all could have the same faulty or out of date drivers, etc.
I'd at least check all the drivers and I wouldn't rely on checking these with the vendors software. You can use Powershell to go through and get the versions on your servers and than start looking up the current versions. Had a similar experience before and we went though this process, found quite a few out of date drivers, updated everything and the problems we had were resolved. I mentioned this at another place so out of curiosity we manually checked some servers - several drivers were out of date. All of these were running the vendor's program to check for necessary driver, firmware updates.
Sue
November 8, 2017 at 11:19 pm
This was removed by the editor as SPAM
November 9, 2017 at 2:51 am
Hi all,
Thank you a lot for your help and Suggestions.
Meanwhile I did a couple of Steps to try to find the root Cause about why the Database got Corrupt in some several servers.
All the Steps since test the IO sub system, look it for out of date Drivers.
Well I cannot have any conclusion about what could be the root cause.
Thank you
November 20, 2017 at 3:20 am
Hi all,
So I am back with the same Topic of the Database issues.
Today I got problems with the Database Integrity.
So can anyone know why the the Indexs got corrupt
Index row in index '$6' (ID 20) does not match any data row. Possible extra or invalid keys for:
with values (Replication Counter = 488896 and Store No_ = '?? and Gate Code = '?? and Entry No_ = 42885) pointing to the data row identified by (Store No_ = '?? and Gate Code = '?? and Entry No_ = 42885).
Any ideas why could be the root cause of this.
Thank you
November 20, 2017 at 4:39 am
GilaMonster - Wednesday, November 8, 2017 3:12 AM99% of corruption is caused by a faulty IO subsystem. Could be anything from filter drivers right down to the disk. There's a lot of moving parts in a modern IO subsystem.
Same as last time. Probably IO subsystem.
That particular one smells like a write cache that isn't properly flushing to permanent storage. Had any unexpected restarts recently? Are all write caches properly battery-backed? Are the batteries still working?
Gail Shaw
Microsoft Certified Master: SQL Server, MVP, M.Sc (Comp Sci)
SQL In The Wild: Discussions on DB performance with occasional diversions into recoverability
November 20, 2017 at 6:40 am
Hi GilaMonster ,
Thank you a lot for your reply But I still cannot know the issues.
I am not saying could not be that but still I tested the Subsystem IO with Diskspd and the SQLIO sim.
The write cache is not enable.
| 11/20/17 | 14:23:03 | 12112 | Display Monitor | DRIVE LEVEL: Read cache enabled = No, Write cache enabled = No |
But I there are any values that I should look with attention.
Thank you for your support
November 20, 2017 at 6:56 am
Performance tests aren't going to do much of anything to test for stability, and Windows not knowing about a write cache doesn't mean there's no write cache.
You need to investigate your IO subsystem. Not run a quick test on it, but look at every level and see what errors, if any, are occurring. Speak with your storage admin, speak with your vendor.
Gail Shaw
Microsoft Certified Master: SQL Server, MVP, M.Sc (Comp Sci)
SQL In The Wild: Discussions on DB performance with occasional diversions into recoverability
Viewing 9 posts - 16 through 23 (of 23 total)
You must be logged in to reply to this topic. Login to reply