No response after a few minutes +blue screen

[Ravenchant]

Hello Sentinel Team,

we have many good experience with your software, but now we have big problems on our new testserver.

After starting HD Sentinel, everything runs well for a few minutes but then, the program shows no response.
A few moments later Windows shows a blue screen and want's to restart.

Components used:

OS:
Windows Server 2012 R2

RAID Controler:
Adaptec ASR 71605 with BBU


Mainboard:
Supermicro H8DGi-F

CPU:
2 x AMD Opteron 6370P

RAM:
64GB

SSD:
First try: system disk:
Samsung 850 pro 256GB

Second try:
Intel DC S3610 200GB

Hard Disks:
12 x HGST HUH728080AL5200 (8TB) as RAID on a Supermicro backplane

We tested many variations:
After we found out that the system drive "Samsung 850 pro" causing many problems, we replaced it with the Intel "Intel DC S3610"

The system becomes unstable always in the moment in which we install HD Sentinal.

I attached the HDSentinel generated DEBUG Report.

Hope you can help.

Greetings
Ravenchant

[hdsentinel]

Thanks for your message and excuse me for the possible issues.

Yes, the best would be to see the developer report (I do not see it attached) as it would help to investigate the situation, examine the "raw" details provided by the Adaptec RAID controller as I suspect the detection of that may be the source.

Not sure which Hard Disk Sentinel version used.
I recommend to check with the latest possible 4.60.10b version released recently:

http://www.hdsentinel.com/beta4/hdsenti ... -ferwq.zip

as it has different updates specifically for Adaptec controllers.
Please check with it - and use Report menu -> Send test report to developer option with that - and let me know if things are better (or at least different in any ways).

[Ravenchant]

Sorry, for the lost attachement.

We use the current testverion "4.60 Pro" for this machine.

I will attache the Debug report as Zip-File and two screenshots with controller information.

Greetings
Ravenchant

[hdsentinel]

Thanks !

Yes, now I see that 4.60 version used. Since that, general compatibility improvements done for Adaptec controllers, so please try as suggested in the previous post, to verify if things are better with that:

I recommend to check with the latest possible 4.60.10b version released recently:

http://www.hdsentinel.com/beta4/hdsenti ... -ferwq.zip

as it has different updates specifically for Adaptec controllers.
Please check with it - and use Report menu -> Send test report to developer option with that - and let me know if things are better (or at least different in any ways).

Thanks !

[Ravenchant]

Hello,

I now have installed version 4.60.10b.
After 10 minutes, the program reported "no response". But it came back to life for one and a half hour.
Then there was finally no more feedback from HD Sentinel.

I've attached the Sentinel error report and the "minidump", which was generated by the Windows Blue Screen.

Greetings
Ravenchant

[hdsentinel]

Thanks.
I checked the dump file but see no references about Hard Disk Sentinel or any of its external modules related to the detection of the RAID controller.

For a quick test, please try to

- completely close Hard Disk Sentinel by File -> Exit
- rename the file deta.dll to deta.xxx
- start Hard Disk Sentinel again.

This way the Adaptec-specific detection will be compltely disabled. The software will not provide any details of hard disks connected to the Adaptec controller. If there are no problems in this case, then we can be sure that the issue is related to the Adaptec controller.
(I thought this - but not sure, as the detection successfully completed and also the minidump shows no references).

If the issue is surely related to the Adaptec controller, I'd recommend to try updating the Adaptec controller driver to latest 7.5 (and maybe update firmware/BIOS if possible) as this may also improve the situation.

[Ravenchant]

Hello,
I've followed your instructions and renamed the deta file on Friday.
The server is running from Friday to Monday without any problems.
Then I reactivate Sentinel and got the old problems.

Of course,updating the Controller-BIOS was one of my first actions.

I tried the original Microsoft Controller-Drivers and the latest Adaptec-Drivers, no difference.

I can not imagine that Adaptec is such an "exotic" manufacturer so that there are generally problems with it.
The Sentinel downloads I received were all "32 bit" versions. Is that the correct version? I mean, there should be no 64bit version?

The mainboard Bios is up to date, too. The penultimate update was 2013, the last 2015.
I can not imagine that there is a general problem with Supermicro motherboards because they're designed for servers.

I also tried almost all BIOS settings to check.
The current settings I have attached as screenshots, is there anything unusual?

Greetings
Ravenchant

[hdsentinel]

Yes, if there were no issues when deta missng, it confirms that the issue is related to accessing the Adaptec controller.

The problem is that from crash dump report, I see no references about the Adaptec controller or any file of Hard Disk Sentinel, as the dump shows

dump_storpor
dump_storahc
cdrom.sys
hwpolicy.sys
sacdrv.sys

files only related to the crash. To be honest, not sure how they can be related....

Also generally interesting that according the report, the Adaptec-specific detection completed without problems.


> I tried the original Microsoft Controller-Drivers and the latest Adaptec-Drivers, no difference.

If possible, please try to create and send developer report again with these.
Not sure, but they may give further ideas as then it is possible to check the differences. Any (even minor) differences can be helpful to investigate the situation.


> I can not imagine that Adaptec is such an "exotic" manufacturer so that there are generally problems with it.

Absolutely not.
Never experinced similar and no other user reported similar with Adaptec controllers.

> The Sentinel downloads I received were all "32 bit" versions. Is that the correct version? I mean, there should be no 64bit version?

No, there is no 64 bit version available - as there is no point of having 64 bit version, since Hard Disk Sentinel does need to utilise high amount of physical memory (more than 4 GB). The 32 bit version is perfect - and runs correctly on high number of systems with Adaptec controllers.

> I can not imagine that there is a general problem with Supermicro motherboards because they're designed for servers.

No, that should be no problem also of course.
Maybe its basic chipset driver can also cause issues - but if I understand correctly, those drivers are also up to date.

I see no problems with the BIOS settings.

[Ravenchant]

Hello Sentinel Team,

I attached the current developer report.
The Bios, all Server 2012 R2 Drivers and the Adaptec driver are up to date.

Greetings
Ravenchant

[hdsentinel]

Thanks, I'm investigating with more details, hopefully will give some ideas about what could cause the troubles.
In the following days, in e-mail, I'll provide an updated version (or a test-tool) to check if there will be any difference.

[hdsentinel]

Just one more thing which may be good to check:

By default, Hard Disk Sentinel detects the status once per every 5 minutes.
According the report, the first detection completes, so it would be nice to verify that the problem occurs
- at the time of the 2nd detection
- any time later, regardless of the detection

If possible, please try this:

1) start Hard Disk Sentinel and when you see the status appears with all information, try to close it completely by File -> Exit.
Perform disk activity (eg. copy/move files or so) and vrify if you experience problems now, after the controller initially accessed.

If you see no problems for longer time (eg. one hour or so), please try starting Hard Disk Sentinel again and check if the problem
occurs immediately (which is the 2nd detect, counting the first detection performed in the previous step) - or just after some time.

Maybe this system simply does not "like" accessing disk status too frequently. You may try to use Configuration -> Advanced options -> Detection Frequency slider to set to detect status once per every hour, just to check if the problem follows the detection cycles.