HP ProLiant MicroServer Owners Club! (Attempting to sell here will result in bans)

I just googled your error without the !!
It seems like flashing the bios is one solution.
https://h30434.www3.hp.com/t5/Deskt...a-SMBios-data-will-not-be/m-p/2580801#M403267

 

Damnit, ok I guess I need to track down the hacked N54L Orom. Maybe I should just try resetting the bios first?

 

I'd just follow the instructions instead of carving my own path here. It's so easy to kill a BIOS.
Look a little earlier for posts on ROMs.

 

I have a N36L with a Remote Access Card.
Running Debian 9.

Debian 9 seems to randomly crash as it did last night so I want a way to monitor and reboot when I can't be at the PC, so looking at hooking the RAC back up.

I don't have the RAC hooked up to network as it's my laundry.
I have a couple TPLink desktop switches but using then they very often crash lights go dim and needed a power cycle.
So one of my questions is what would the best way to use the remote access card where I have 1 network cable to the laundry.

Was looking at second hand 5 Port HP switches 1410 I have a 24 port running great as my main switch.

Also with the remote access card I think I can see Debian throwing up messages about it on boot.
So will need to look into that.

I'm just wanting to know three best way to hook it back up that's all.

 

Sorry you've had stability issues with the TP Link stuff. I only have one 5 port Gigabit switch and its TP Link. It's in my workshop hooked up permanently, but doesn't always have any traffic across it as I barely ever work on desktops anymore, so I can't reflect on it's stability.

I can recommend the DLINK rack sized switches.
https://www.ebay.com.au/itm/DLink-DES-1024D-24-Port-Desktop-Switch-D-Link-Free-Shipping/183787594363
I've had my 16 port running 5 years without a hiccup.
If space is an issue maybe you'd like to put in a 30cm deep 2RU or bigger rack up high in the laundry.
https://www.ebay.com.au/itm/2RU-19-Wall-Mount-Rack-Frame-300mm/392253592365
The install with DLINK 24 port at my brothers was a 4RU rack so a little small to house the Microserver (just right for a ReadyNAS), so the server could sit on top.

 

/var/log/syslog

or
/var/log/kernel

Look in kernel log first for anomoly near the reboot time, if no joy look in syslog might be a clue there.

If your lucky you might have a coredump.

Best to figure out why its randomly crashing.

 

I didn't mention I'm renting so my main HP switch is in the loungeroom in a 'LackRack', my modem router is at the door, the switch is a unmanaged 1410 24g port with the cables going around the edge and under the door edging.

So there's 1 line to the laundry area.

May pick up the HP 5 Port 1410 5g port as I haven't had any issues with the 24 port I use as my main been rock solid through anything.

Was just wondering if there was a better way than having a separate switch for the microserver and RAC.
Could mby look at a Lan to WiFi adapter.

Yea I'll look at the logs today and see what happened.

 

No coredump and what just looks like standard kernel logs.

I'll keep a eye on it.

 

Tried resetting bios (is it just "load optimal defaults") but no dice fixing the SM Bios error.

The system still works wonderfully, but it's not a good look when I go to sell it.
That being said, has anyone else in the thread inserted the special 16GB memory into it? I managed to find those modules which give it 16GB and maybe it's that? (I never have posted it attached to a monitor with the 16GB, lol! I just bought it, took it out of the cupboard, put the ram in and put it back, booted up FreeNAS fine)

 

Flashed this bios.
https://cloudrun.co.uk/server/bios-for-hp-proliant-microserver-g7-n54l/

Erased the bios
Pulled the battery and held down power
Still getting the error, it says press F1 to continue, but it just keeps working. I'm wondering if a heap of people are getting this?

302 not enough space in runtime area SMBIOS data will not be available? Apparently it might be an event log or something?

 

I just remembered there are these cheap Chinese mining pc usb watchdogs.
You can use a app and turn the PC on/off and when the PC crashes I'tll automatically restart it.
I decided to get a cheap one today and will play around with it on a second system as I was thinking of using it in the HTPC for remote power from my phone. I am sure there are other devices to do that.

If it works I was thinking of hooking it up to the microserver as a automatic restart tool.
From what I can see they monitor a signal over the USB port and if it stops will restart the PC.

The N36L and others HP etc all have a special non standard connector to the board for the power reset led's etc.
So I would need to cut the power/reboot switch and add this inline and use the internal USB to power it as it has a power usb connector then a 2 pin for the power button.

Might be something for the rest of you to consider etc I may make a seperate thread about it in the electronics or modding forum will need to see whats the most relevant for it.

 

Isnt there a setting in iLo to have the machine restart on power failure ?

I know when ive had power cuts my server is always running afterwards, although complaining about it

 

That's a common Mobo option your mentioning.
After AC loss.
Turn on, Keep off, Continue as was before loss.

What I'm talking about is a device that monitors a active signal so if the system locks up it'll reboot it automatically.
These devices are known as watchdogs and are used in systems dedicated to mining etc.
I'm sure proper servers have a heartbeat monitor that does much the same.

The Chinese ones are only $10 or so each and can use a Android app eWeLink to control the power also.

 

Hi Everyone,
I thought I'd report on a fairly serious issue I'm having with a MicroServer Gen 8. This is the second time I've now seen the problem. Essentially I see things like this in the Linux /var/log/kern.log file:

Code:

Jul  3 06:26:10 server kernel: [126727.440088] ata5: SATA link up 1.5 Gbps (SStatus 113 SControl 310)
Jul  3 06:26:10 server kernel: [126727.440625] ata5.00: supports DRM functions and may not be fully accessible
Jul  3 06:26:10 server kernel: [126727.441693] ata5.00: supports DRM functions and may not be fully accessible
Jul  3 06:26:10 server kernel: [126727.441978] ata5.00: configured for UDMA/33
Jul  3 06:26:10 server kernel: [126727.441983] ata5.00: retrying FLUSH 0xea Emask 0x10
Jul  3 06:26:10 server kernel: [126727.442127] ata5: EH complete
Jul  3 06:26:19 server kernel: [126736.254218] ata5: exception Emask 0x10 SAct 0x0 SErr 0x4090000 action 0xe frozen
Jul  3 06:26:19 server kernel: [126736.267670] ata5: irq_stat 0x00400040, connection status changed
Jul  3 06:26:19 server kernel: [126736.281128] ata5: SError: { PHYRdyChg 10B8B DevExch }
Jul  3 06:26:19 server kernel: [126736.294392] ata5: hard resetting link
Jul  3 06:26:29 server kernel: [126746.317515] ata5: softreset failed (device not ready)
Jul  3 06:26:29 server kernel: [126746.330686] ata5: hard resetting link
Jul  3 06:26:30 server kernel: [126747.225589] ata5: SATA link up 1.5 Gbps (SStatus 113 SControl 310)
Jul  3 06:26:30 server kernel: [126747.226397] ata5.00: supports DRM functions and may not be fully accessible
Jul  3 06:26:30 server kernel: [126747.227456] ata5.00: supports DRM functions and may not be fully accessible
Jul  3 06:26:30 server kernel: [126747.227739] ata5.00: configured for UDMA/33
Jul  3 06:26:30 server kernel: [126747.227744] ata5: EH complete

When I first noticed these messages, it was on ports ata1 > 2 and from time to time, on 3 & 4. At first you might assume the drives are failing, but after a little testing I didn't believe that was the case. The result of these errors were two fold; one everything would come to a screaming halt and two, data corruption - very serious corruption. When I first had this issue back in May 2018, I replaced the power supply (200W not the 150 version) under warranty and the issues went away.

My research at that time suggested power or SATA cables might be at fault, but I ruled that out as well. I've now had this issue pop up again after a few weeks so I replaced the PSU once again. I removed the third and fourth drive as well, but straight away noticed the errors did not reoccur on ata1 or ata2.

Fast forward a week later and I'm now seeing it on port 5 which I've connected a Samsung SSD to. Because the SSD runs the OS, when it misbehaves the whole system crashes to the point where its almost impossible to restart. On occasions you'll also be prompted at boot to run fsck to repair errors on that partition. HP iLO and ADH logs of course show nothing of any real consequence. It would appear that this becomes an issue when you've placed any basic sort of load on the system. I have been able to recreate the problem on occasions with using DD to put together a 1GB file, and more recently when I set up a VM on the SSD and put a bit of load on it.

I'd certainly welcome the feedback of those more proficient in Linux than I.

 

Funny you should mention this; I had to do the same to prevent errors on my Linux system like this: failed to get NCQ Send/Recv Log Emask 0x1.
I believe this issue was fixed in a kernel update some time ago, but I was too afraid to update the firmware of my SSD.

 

What graphics card is everyone running these days?

Is the HD6450 still suitable or would the GT1030 be a worthy upgrade if it fits? or is there something better?

 

Does anyone have humungous (as in 10tb drives) in their microserver? Mine has 4tb, it's fine with those as long as you format them in another system first. Can it cope with huge?

 

I had 10tb's in a gen 8 microserver no worries and ive seen 6's and 8's in the the n54l's running fine.

 

Thanks, Mutman, it's Thunderbirds GO!

 

Only issue i can think of is a hard limit of 16tb per volume under 32bit linux.
Which microserver you running ?