Server keeps stopping moving of large files to it 1

[gavm99]

Hi all,

This is a really strange one.

My server is running 100% fine in terms of services, email, internet access and there are no errors in the event logs.

The problem I have is when I want to move large files from my client machine to a share on the server. By large files, I mean for example 4GB. It starts to move the file and all looks to be fine but then after a short period of time it will stop and then it will come up with the prompt telling me the file already exisis and asks me if I want to replace it, even though the file doesn't actually exist there. It is like it has started to move the file again. If I click Yes it just does the same thing over and over.

Smaller files are fine.

I have tried from a different machine and the same happens.

Any ideas?

Thanks.

----------------------------------------
"Nobody cares how it works, as long as it works

 

[wahnula]

Are the clients wireless or wired? Any errors in the server EventLog regarding the LAN NIC?

Tony

Users helping Users...

 

[gavm99]

Both clients are wired at 1Gbps, I will check the client event logs are report back.

----------------------------------------
"Nobody cares how it works, as long as it works

 

[wahnula]

Oftentimes a hardware event will be a yellow-triangle Warning, not the bloody red X associated with an Error. I had a problem with my Broadcom LAN NIC that would be intermittent for no apparent reason, it came down to a carelessly coiled CAT 6 loop. The Warnings were on the server and the client, but were easy to miss because they were not Errors and thus not reported in SBS Monitoring & Reporting.

To eliminate the network & wiring from the equation, you could try a large file transfer from a flash drive or a DVD connected directly to the server.

If that works fine, eliminate the switch/router/cabling and connect a client directly to the server. If you have a laptop, you could do it in the server room with a fresh cable for a baseline.


Tony

Users helping Users...

 

[gavm99]

Hi,

There are no errors in the event logs so no clues there.

This morning I tried the same move of data and this time it worked ok?!

I have rebooted the switch and router for good measure but my gut feeling is that this is done to disk throughput ( 2 x SATA disks, not RAID ) as this morning when the move did work it did also upset my connectio to Exchange saying it was unavaliable during the data move and then coming back online afterwards.

I am going to try a quick test and move the server page file to a dedicated disk to see if this helps matters.

Thanks.

----------------------------------------
"Nobody cares how it works, as long as it works

 

[wahnula]

Is there a real-time virus scanner on the server? Just a WAG...

Tony

Users helping Users...

 

[gavm99]

Hi,

No it is not an AV issue as I have uninstalled this to be certain.

I do have some more finding though:

I moved the page file to a seperate dedicated and this essentially has made no difference.

I have connect a USB direct to server and attempted to copy a 4GB file to both the C partition and the D partition and again the server just appears to lock and not actual perform the move. The server is pingable when this lock occurs but you can't contact Exchange or access file shares.

Under inspection of the event log I have now found this which might be clue:

System Log, Service Control Manager, 7011

Timeout (30000 milliseconds) waiting for a transaction response from the NtFrs service.

Can anyone help with this error? I am sure this is linked to my problem.

Thanks.


----------------------------------------
"Nobody cares how it works, as long as it works

 

[ShackDaddy]

I don't know if that error is part of the cause. Since it's a replication error, it's probably because the replication service is having the same issues accessing the disk that the Exchange service is.

Let me make sure I understood the config:

You have two SATA disks. Are they in a mirror with C: and D: as partitions on that mirror? Or do C: and D: each have their own physical disk? And what's the speed on those drives? 10k or 7200?

When you work with SATA, the controller is the potential bottleneck, since it has to handle all data transfers regardless of the number of disks. Perhaps your controller firmware/driver needs an update.

Dave Shackelford
Shackelford Consulting

 

[gavm99]

Hi,

Yes I have 2 x SATA disks which are mirrored. C and D both exist on the same disk.

I have done extensive testing over the last 24 hours and I have found that:

1) It is definatly not the server's LAN card, the switch or router. I know this as I have changed all 3 and I still get the same problem.

2) It is not any other server hardware issue as I used a spare disk and did a basic SBS install on that and everything works fine, file copies to it are rapid and the server responds fine.

3) It is not 1 of the disks in the server as I mirrored my basic install to one of the actual server disks and this worked fine. I am yet to test the other disk on its own but I am certain from the testing about this is done to a software issue.

4) As stated on my last point I am 90% sure is down to software.

Overnight I am going to create a virtual machine of my server (as I did it working as my mail server tomorrow) and this way I can test this as much as I like without stopping any business activity.

Please put forward your software suggestions!!!!!

Thanks.

----------------------------------------
"Nobody cares how it works, as long as it works

 

[gavm99]

Mmmmmmmmmmmmmm this is getting confusing now, is it hardware (as in 1 of the disks) or software?

Before I went to bed last night I set the mirror going again. I would expect this to take a few hours only to complete as C parition is 25GB and D partition is 208GB.

However I have come to check it this morning and the C partition has completed but the D partition is only 65% done. This isn't right!!

When I mirrored my test install mentioned earlier in this posting it took a few minutes to mirror 20GB.

Could software be causing this problem?

I can hear the disks now, they go mad for a few minutes, stop for a few minutes, and then go mad again, over and over.

Confused!!!

Oh, and I also got this alert via email:

The following disk has low idle time, which may cause slow response time when reading or writing files to the disk.

Disk: 1

Review the Disk Transfers/sec and % Idle Time counters for the PhysicalDisk performance object. If the Disk Transfers/sec counter is consistently below 150 while the % Idle Time counter remains very low (close to 0), there may be a problem with the disk driver or hardware.

If the review shows that the disk is functioning properly, use Task Manager to determine which processes are causing the majority of the disk activity. You can attempt to correct the problem by stopping and then restarting those processes.

You can disable this alert or change its threshold by using the Change Alert Notifications task in the Server Management Monitoring and Reporting taskpad.


----------------------------------------
"Nobody cares how it works, as long as it works

 

[gavm99]

Yet more testing, more progress (sort of).

Ok so after much messing out I have the copying of files problem sorted at the moment. I have the server running on 1 disk (I know, I know) and all is well in terms of file copying and moving.

However I have noticed that the server can map a drive to a networked USB drive now. The USB drive is ok, other machines see it fine. The server says that it does not have access or it is because of performance.

I have tried reinstalling SP2, I have ran a repair of SBS and now I am just awaiting the restore of the system state from 2 weeks ago.

This is looking more and more like a reload..........

----------------------------------------
"Nobody cares how it works, as long as it works

 

[wahnula]

Before I went to bed last night I set the mirror going again. I would expect this to take a few hours only to complete as C parition is 25GB and D partition is 208GB....Could software be causing this problem?

I can hear the disks now, they go mad for a few minutes, stop for a few minutes, and then go mad again, over and over.

The only software that I am aware of that could be causing this problem is the software running your RAID controller...is it integrated into the motherboard or an add-in card? What is the make and model of the RAID controller?

If you mirror the known good disk to the second disk, then break the array and run SBS from the second disk only, what happens? I am not liking either that RAID controller or the second disk.

I have had nothing but problems with motherboard-based RAID controllers. I prefer 3ware RAID cards which are more truly hardware-based RAID solutions.

Tony

Users helping Users...

 

[gavm99]

Hi,

Still working on this but in summary:

The disks connect direct to the motherboard, there is no hardware RAID.

What I have found that (yes I have changed my mind) that this is down to a dodgy disk.

To keep it simple and explain things better we will call them WD disk and HP disk.

WD disk is fine on its own, rapid transfer speeds and does not lock up server at all.

HP disk seems dodgy, causes server to lock up and took 14 hours to mirror.

I am going to do a final test with the HP disk though, I will do a low level format and then mirror the healthy WD disk to it and see what I have afterwards.

The USB issue where the server couldn't see it is a red herring, the same happens with a clean install.

Right now I am restoring data to my newly created data partition on the WD disk (I recreated the data partition to be safe).

Once this is done I will start work on the HP disk.

Thanks.

----------------------------------------
"Nobody cares how it works, as long as it works

 

[ShackDaddy]

Glad you've been able to narrow it down. I had a strong feeling it was hardware, and I was worried about how much time you'd spend looking for a software problem.

Dave Shackelford
Shackelford Consulting

 

[gavm99]

Nailed it!!!

I reformatted the HP disk. I then mirrored to it but only the C partition to save time. This completed (taking too long for my liking) and now when I try my copy tests, it once again fails!!!

The source of all evil is this dodgy HP disk!!

Treating myself to an early night, warranty call tomorrow.

Thanks all.

----------------------------------------
"Nobody cares how it works, as long as it works

 

[wahnula]

Congratulations gavm99!!! A star to you for your persistence and keeping us informed of your process/progress.

I know you used "HP disk" and "WD disk" as a means of identifying the disks to us, but I hope they are actually the same. RAID always works best with drives of the same make, model, speed, and capacity. If they are actually physically different disks I would consider buying another "WD disk" and using the new "HP disk" for backups or as a utility disk.

I have used everything but the kitchen sink for RAID arrays and they have all worked, but this is a server and best practices should be followed.

Well done! Enjoy your hard-earned break.

Tony

Users helping Users...

 

[gavm99]

Hi all,

Just as a final wrap up to this:

The new disk has arrived from HP and the 2 disks are same as recommended. Having rebuilt the mirror in a matter on minutes all is now working great.

Thanks.

Gavin Moorhouse

http://www.lucidcomputersolutions.co.uk