report email

[VGER]

Hello,

HD Sentinel has started to fail sending its daily report-email on one of my systems. There is an error popup in the lower right corner (which keeps popping up again and again apparently until the mail can be sent, which is never):

Status e-mail sending error.
Out of memory while expanding memory stream

If I quit HD-Sentinel and start it again, the email can be sent immidiately, and "mail test" also succeedes. But the next day when the report mail is due, it fails again.

It feels like this has started after I swapped the system drive on the machine from SATA-SSD to NVME-SSD. But why would that cause problems with memory, and why can the mail be sent just fine after HDS is restarted?

Regards

[hdsentinel]

Hello,

Thanks for your message and sorry, but I'm almost sure that it is independent from Hard Disk Sentinel and generally should be independent from the SSD of course. Not sure, but somehow /something may prevent normal operation. Never ecountered any similar and no user ever reported similar.

Not sure, but it is possible that some file(s) could be damaged. I'd surely try a complete "fresh" clean installation by uninstalling, deleting the previous folder where originally installed ( C:\Program Files (x86)\Hard Disk Sentinel by default) to delete all previous statistics, status, settings and then install again to verify if you encounter similar after a clean (re)installation.

If you encounter similar, I suggest to please use Report menu -> Send test report to developer option: checking the actual situation can always give some ideas, thoughts.

[VGER]

Actually, the situation got worse: HDS is currently not usable on this system. Apart from the email issue, HDS hangs ("not responding") and more erratic behaviour ("zombie windows" in the taskbar etc.)

As I mentioned, the problems started right after I exchanged the system drive. Apart from having to clone the drive (using Clonezilla), I just found that since there is a system service running named "SamsungMagicianSVC". (The Samsung Software is installed, but not running full time. I hadn't noticed the on-by-default service before.)

I have now disabled the Samsung service and restarted HDS. So far it seems to be ok, but I need to watch it a while to be sure.

Do you know of any issues with this Samsung service? (Installed is "Samsung Magician" v7.2.1.980.)

Regards

[hdsentinel]

When HDS is "not responding" it means that it is waiting for a device (disk drive / controller or so) to respond. It does not hang - just the device is not responding as expected. Yes, in this case HDS seems "not responding" as the device is not responding - and HDS is just waiting.

No, there are absolutely no issues with Samsung Magician. Never encontered any issue and no user ever reported any issue.
I'm sure it is not the background, but something else (one or more of your drives may not responding as should).

[VGER]

There are a number of aspects here that make this issue very confusing. I'll try to separate them:


1.) report email not sent:

Yesterday, I restarted HDS and tried "send test mail". The email was sent, HDS was fully usable (browsing disks, going through menus, config dialog.)

This morning I tried again: Same, all was well. Mail test successful.

At 1215 HDS is supposed to send its daily report. This email was not sent.

I had Wireshark running, monitoring the smtp session. The test send (before noon) went perfectly fine. At 1215, HDS tried to send, but was not successful. The session looks like this:

S: 220 server.tld.de blablabla
C: EHLO
S: 250 Hello <capabilities>
C: AUTH LOGIN
S: 334 <challenge>
C: QUIT
S: 334 <challenge>

... so HDS just aborts the handshake ("C: QUIT") instead of responding to the server challenge as it did before (with the successful test mail.)

This handshake attempt repeats at first every 1.5 minutes, since then every five minutes. It is still going on now (past 1500).

On every attempt, HDS shows its popup with the error message cited in my first post.

The SMTP server is fine, because other clients can successfully transmit during all this time.


2.) HDS not responding:

I just found there were two HDS processes running on the system, one with fairly ridiculous memory allocation (1.9 GB). This seems to be what made HDS "hang" - apparently it was fighting itself for access to the hardware.

I'm now looking into where that second process might have come from. For the moment, I've deactivated all autostart options (Windows Autostart folder, Scheduled Task Management, services). For now, I'm starting HDS manually only, and the second process has not reappeared.


Regards

[hdsentinel]

Thanks for the details. Both issues sound really interesting, weird - as generally (of course) these should never happen.

1) Maybe the e-mail contents can be somehow in the background.
What happens if you try to open Configuration -> Message settings page and try to
- change the report type to different
- change/disable some of the report components

and also use the manual sending options: Report menu -> Send TXT report or Send HTML report

just to verify if things are different and track down if some report option(s) may affect the results.


2) Generally 2 instances of Hard Disk Sentinel is not possible at all: when Hard Disk Sentinel is running and you start again, the 2nd one detects that it is already running and closes automatically. So I really do not understand how you saw two different instances.


I still suspect that something in your installation could be damaged/corrupted, so I'd surely try what I wrote: a complete "fresh" clean installation by uninstalling, deleting the previous folder where originally installed ( C:\Program Files (x86)\Hard Disk Sentinel by default) to delete all previous statistics, status, settings and then install again to verify if you encounter similar after a clean (re)installation.

I'd surely install the latest 6.10 version (not sure if you checked that).

[VGER]

Hello,

after some days of observation, it seems to me your original hunch was correct: I've now disabled all auto-starts (no more "dual" processes so far) and I've disabled email sending completly (no more popups about failed messages.)

However, the "hangs" are still there: HDS is running minimized, but can't be opened by clicking on its tray icon. You CAN right-click on the tray icon and open settings, but only once (a second time the dialog won't open) and not all pages are accessable in the dialog once opened: The "first few" pages are ok, then something happens and the right hand side isn't properly redrawn any more. It doesn't seem to matter which pages you select to make that happen.

Which brings me back to when it all started: After I had replaced the system drive SATA-SSD (Samsung 840 pro) for an NVME-SSD (Samsung 980 pro). The NVME-drive is connected via an extension card directly on the PCIe-slot (the mainboard connects its M2-sockets to a single PCIe line only, thus the slot-card.) Might there even be some BIOS problem?

I will send you a report from HDS directly in addition to this thread.

PS: Yes, I've seen your suggestion for a full reinstall of HDS. But I explicitly do not want to loose the database of existing and past drives. I have since reinstalled "over" the existing version, and there also was a recent version update (now running v6.10) which I had HDS perform automatically. So at least the binaries should be fine.

PS2: The "hanging" HDS process again is shown consuming 1.9 GB of RAM allocation "private bytes".

Regards

[hdsentinel]

Thanks for the info.

I'm sorry, but I'm almost sure something is damaged or not working correctly in your system

I doube checked the Hard Disk Sentinel report, I see the Avago card with the hard disks and the PCIe SSD too.
Generally there should be no problem with the NVMe SSD or the hard disks, the controller or other storage related devices. If Windows can boot with no problems then I assume there should be no issue with the BIOS too (but you may check if there is a BIOS update, just to be sure).

However, from the developer report, I see many entries when the software can't work as expected as Windows reported

System Error. Code: 8. Für diesen Befehl ist nicht genügend Speicher verfügbar
( Not enough memory resources are available to process this command. )

This is usually not related to the system memory, Hard Disk Sentinel can never use of your complete memory of course - but the fact that it can't work properly causes that it can't free up what it may use - which result the high memory use which can simply never happen in normal situations.

Error 8 is usually related to some bad driver, a non-correctly working hardware, corruption of the registry or some Windows components
(you may find several results if you search for "Windows system error code 8").

First I thought the binary files (and/or the status/statistics/logs data files) of Hard Disk Sentinel may be damaged, but I suspect this is not the case.

I understand and agree that you do not perform a clean (re)install (I'd probably feel the same).

Maybe you can try to
- completely close Hard Disk Sentinel by File -> Exit
- try to unpack the latest portable version
https://www.harddisksentinel.com/hdsentinel_pro_portable.zip
to a completely new, empty folder, eg. C:\HDSPortable
(DO NOT in the folder where the current version installed)

and launch HDSentinel.EXE from that to have a clean start: it works independently from the installed version, will not use any of the existing settings/statistics). Just to observe if things are same for some hours/days/weeks - or different.

[VGER]

I may have found something:

Apparently since I swapped in the NVME drive, I'm getting a massive amount of errors in the system log:

Protokollname: System
Quelle: Microsoft-Windows-WHEA-Logger
Datum: 17.03.2023 13:04:44
Ereignis-ID: 17
Aufgabenkategorie:Keine
Ebene: Warnung
Schlüsselwörter:
Benutzer: Lokaler Dienst
Computer: <machine name>
Beschreibung:
Behobener Hardwarefehler.

Komponente: Stammport für PCI-Express
Fehlerquelle: Erweiterte Fehlerberichterstattung (PCI-Express)

Bus/Gerät/Funktion: 0x0:0x1:0x0
Lieferanten-ID/Geräte-ID: 0x8086:0x1901
Klassencode: 0x30400

This is the port the NVME is connected to. Apparently, it has something to do with incompatibilities of PCIe power states, though I haven't been able to confirm that yet - the suggested workarounds (disable power save in Windows Energy settings) don't seem to have any effect. I'll try modifying the BIOS settings at the next reboot opportunity, maybe this will make a difference.

Now finally all of this is starting to make sense: Apparently, HDS tries to check the Event Log before it sends its report email. If there are lots of entries (literally thousands in my case!), something overflows and not only does the email send fail, but this buffer keeps hanging around somewhere and gets larger and larger over time, until HDS first freezes and then crashes. That's the ridiculous RAM draw and the hangs I've observed.

For now, I have disabled "monitor system log" in HDS. If the theory is correct, then HDS should now work normally again, even if the root problem (the numerous log entries) is not yet fixed.

Regards

[VGER]

Ok, theory confirmed.

HDS is now working normally again on the machine, no more hangs and reports are again sent by email. RAM usage of the task is normal again as well.

For those interested:

The flood of event log entries appear to be a problem with the mainboard (Asus P10S WS in my case): As soon as a device is plugged into one of the PCIe-slots connected directly to the CPU, the events are generated. Plug the same device into the PCIe-slots connected to the chipset, and there are no more events. Of course, then all data has to pass the DMI link between chipset and CPU, which may act as a bottleneck.

Since I wasn't able to fix the root cause for the events (they seem to be more of a nuisance than an actual error condition), I just disabled them. Not exactly a joice solution, but so far it seems to work as intended.

Regards