mats42
-
Posts
26 -
Joined
-
Last visited
Posts posted by mats42
-
-
well. what I did was shutdown the box from the gui, replaced the usb memory with one with 1.02
booted the system, klick on the upgrade to 6.1.1. it installed and did an automtic reboot.
after that it newer booted again, I tried all three bootalternatives - no boot and not detectable by find either. Even tried with reimaging the usb device just for saftey.
Can it be something else at the same time - maybee.
Is it a big deal that data was lost? nope - it took a few mins to recreate the backup job and the box fixed the rest during the night.
The point is that you should have backup or be prepared to loose data when playing with alpha stuff. I was and I'm happy with my upgrade
-
tried to update 6.02 -> 6.1.1 on bare metall
Did it work - No
Did I loose all data - Yes
Do I blame somebody for it - No
Did I have to reinstall - Yes
Am I happy - Yes
It is an Alpha loader and my box is only used as a backup for the true server so after reinstall, I just configured the backup jobb again. It actually took longer time to clean the disks for the new install
Updating, upgrading reinstallling does not cause loss of data. System and data are in two very distinct partitions.
having the partion table go corrupt does ......
-
tried to update 6.02 -> 6.1.1 on bare metall
Did it work - No
Did I loose all data - Yes
Do I blame somebody for it - No
Did I have to reinstall - Yes
Am I happy - Yes
It is an Alpha loader and my box is only used as a backup for the true server so after reinstall, I just configured the backup jobb again. It actually took longer time to clean the disks for the new install
-
Hello quicknick,
it would be really great if you can implement some drivers for:
Realtek RTL 8139
With this it would be possibile to get DSM 6.0.2 working under XenServer 7.
At the moment the last working version of XPE is 5.0
Many thanks.
Bye
gmop
I don't do Xen myself but I found this for you
-
hi,
DSM 6.0.2-8451 Update 5 works with this loader ?
thx
YES
-
Hi mats42,
The sensor is just a cable connecting to the board. I just disconnected that cable. Its possible that your server might be different to me. On my ML10v2 the ambient/intake sensor is sitting at the front of the server, and the connector is at the bottom of the board. If you track the sensor cable , you will know where it is.
Thanks.
I missed that you had an ML server.
I will dig deeper and see if i can find where the sensor is on my microserver
Dead Easy on the Microserver to. The Tempsensor connector is located on the inside of the powerconnector on the mainboard. It has a brown loop on it.
And removing it gives a reported 6% fanspeed on the microserver. Of course I take no responsibilty what so ever for this
-
Hi mats42,
The sensor is just a cable connecting to the board. I just disconnected that cable. Its possible that your server might be different to me. On my ML10v2 the ambient/intake sensor is sitting at the front of the server, and the connector is at the bottom of the board. If you track the sensor cable , you will know where it is.
Thanks.
I missed that you had an ML server.
I will dig deeper and see if i can find where the sensor is on my microserver
-
Hi,
You are I think exactly correct. My HP ML10v2 runs on 6% all day when the temp here these days is between 22C to 37C during the day, when running the HP ESXi u2 with VMs on it. Only if the xpenology Vm is doing the plex transcoding or something, then the fan will go a bit higher may me 15% or 20%. But, when running baremetal it runs at 18% to start off the day at about 22C, and then keeps on going up during the day to 60% at about 31+C. Which is very annoying. So something is really missing there, and I think there are the drivers issue. I have successfully included the AMS module that I found in another forum, so I can now see the AMS data , but still the fan is issue is there.
Thanks.
Look here for how to fake the fan signal, etc..
http://homeservershow.com/forums/index.php?/topic/7294-faking-the-fan-signal/page-1
Ciao
ema
Hi,
Thanks for the link, but its a two year old link and lots has changed with the Bios and iLO fw updates. Its not the same issue as well. What we need is the drivers from HP ESXi u2 injected into xpenology like the AMS module i've managed to work. Just don't know how/where the iLO drivers. The only solution that I found that works is to disconnect the front temp sensor and makes the iLO think that its running at 0C and the fan stays at 6%. Of-course if the cpu load/temp goes up fan spins higher. There seems to be no ill-effect of disconnecting the front intake temp sensor.
By the way, DSM6 runs fine as a VM in HP ESXi u2 with no fan issues. Fan stays at 6% at any ambient temp for me. The only thing is that the HDDs wont spindown/sleep as they are running in RDM mode from the HP B120i controller.
Can you tell me how you did disconnect that sensor?
-
Hi,
You are I think exactly correct. My HP ML10v2 runs on 6% all day when the temp here these days is between 22C to 37C during the day, when running the HP ESXi u2 with VMs on it. Only if the xpenology Vm is doing the plex transcoding or something, then the fan will go a bit higher may me 15% or 20%. But, when running baremetal it runs at 18% to start off the day at about 22C, and then keeps on going up during the day to 60% at about 31+C. Which is very annoying. So something is really missing there, and I think there are the drivers issue. I have successfully included the AMS module that I found in another forum, so I can now see the AMS data , but still the fan is issue is there.
Thanks.
Look here for how to fake the fan signal, etc..
http://homeservershow.com/forums/index.php?/topic/7294-faking-the-fan-signal/page-1
Ciao
ema
No, that's more about fooling the termal protection (ie killing of the rotation detect protection) which might be a bad but unavoidable workaround if we can't fix the real problem
If you install Windows, Vmware, Redhat Enterprise, Suse, Fedora or Centos with propper drivers the fan will run at 6%.
If you instead install Xpenology the fan will be at 12-15% on exactly the same hardware under the exact same conditions.
If you dig deep into it the problem is that the ILO reads temps from all drives as a base value for it's cooling settings. In AHCI mode or in raid mode without the propper driver loaded into the running OS, the ILO can't get the temp values and therefore has to use (higher) failsafe values for the fan speed. By loading the right driver - ILO now can read the temp Values and adjust the fan speed properly
Therefore the fix is to make sure the ILO can get temps under XPenology and only if and when we know that it can't be done - then it's time for "dirty" HW hacks - prefferbly one that just adjusts the fan speed a little lower so that the protection system still works as intended.
-
I don't know about that specific drive but the controller will do 8GB drives at least. See http://homeservershow.com/forums/index. ... n-8/page-2
-
1) 2 nd NE555 gives a higher RPM pulse based signal to the ILO. So it thinks the fan is @15% actually the other NE555 is running at a noise satisfying level you control yourself,
2) Sure you are right but come on how many times did you encounter this?
3) I'm sure that when you set the rpms yourself the temperature will be at acceptable levels in all conditions. My i7 does have fixed values my case has fixed levels never had problems with temps.
4) HW hacling is not the case. You feed 2 NE555's with the original 4Pin Fan output of the motherboard (ilo) and distribute a new connection to the fan and one lead back to the pin whioch measures the RPM.
5) because you can do it yourself. The SW issue wont be fixed by HP I think.
1. You miss the point. There is no RPM signal in the connector so you can't send one. The ILO requires a different type of signal (rotation locked/tach). If you try to send a RPM signal on that wire the ILO will see the fan as faulty and shut down the system directly.
2. Not often but have I seen it happend on HP servers -yes and why take the risk when you don't have to
3. Not on a microserver. At 10% system load with 20C ambient temp (calm winter day) about 6-8% is enough. At 90% cpu load it will need at least 15-20% and the noice difference is huge so therefore a temp controlled solution is the best
4. Well, two small problems - there is no 3 or 4 pin fan connector in a microserver .......... Only a HP only six pin and there is no rpm pin in that since the ILO doesnt messure RPM. It messures temp and if the temp isn't right it orders the fan to higher speeds. The ILO will also order higher fan speeds if it can't read drive temps - Values it tries to read from the missing driver
5. Since the "problem" is that Xpenology doesn't contain the driver HP do provide - HP can't fix it but we may be able to if we include the ILO amd AMS drivers. Hence the request for including them. The ams driver was hacked into 5.X-somwthing (I read about it earlier)
-
coI have been looking into the noice/fan issue of the Microserver G8 (Yes - it still exists with the latest bios 15% vs 6%).One thing that i think could be interesting to test is to include the Binary only driver for the Raid wiController (https://launchpad.net/~hp-iss-team/+arc ... hp-storage).
We know that the fan in the G8 is controlled through ILO and that the ILO needs this driver to be able to get drive temp data.
Since the Controller do support creating four Raid0 arrays of one disk each we still could let the OS do the true raid work.
Just my 2c
i would just create a new pwm controller with 2 NE555's 1 to control the actual fan itself the other for making the ilo happy. You can slowdown the fan to let say 4 or 5 % lowest possible pwm setting to let the fan rotating and the other setting pwn pulses which enters the ilo. its about 1 or 2 euro on components.something like this http://i.imgur.com/Hw1zBKO.png
I sed a few problems with that approach.
1. the fan doesn't give an rpm signal back so you can't send a pulse based signal to the ilo. It needs a tach/rotation locked signal
2. If the fan fails, a faked signal would allow the server to overheat
3. If the temp in the server rises a fixed PWM signal would allow the server to overheat
4. Hw hacking would void warranry
5. Why do a hw workaround on what we know is a software problem instead of fixing the SW? we know that the fan will run at correct speed with the right sw
Sorry
-
coI have been looking into the noice/fan issue of the Microserver G8 (Yes - it still exists with the latest bios 15% vs 6%).One thing that i think could be interesting to test is to include the Binary only driver for the Raid wiController (https://launchpad.net/~hp-iss-team/+arc ... hp-storage).
We know that the fan in the G8 is controlled through ILO and that the ILO needs this driver to be able to get drive temp data.
Since the Controller do support creating four Raid0 arrays of one disk each we still could let the OS do the true raid work.
Just my 2c
i would just create a new pwm controller with 2 NE555's 1 to control the actual fan itself the other for making the ilo happy. You can slowdown the fan to let say 4 or 5 % lowest possible pwm setting to let the fan rotating and the other setting pwn pulses which enters the ilo. its about 1 or 2 euro on components.something like this http://i.imgur.com/Hw1zBKO.png
I sed a few problems with that approach.
1. the fan doesn't give an rpm signal back so you can't send a pulse based signal to the ilo. It needs a tach/rotation locked signal
2. If the fan fails, a faked signal would allow the server to overheat
3. If the temp in the server rises a fixed PWM signal would allow the server to overheat
4. Hw hacking would void warranry
5. Why do a hw workaround on what we know is a software problem instead of fixing the SW? we know that the fan will run at correct speed with the right sw
Sorry
-
mine is stock 1610. No extra cards and four disks. Runs on 12-15% in ahci with XPenology. In Redhat with raidmode it goes lower.
I like silent things and I can't see a reason why we shouldn't be able to get that under Xpenology
-
First off all:
Thank you. Great to see a new loader for DSM 6
Second:
I'm interessted at the HP Gen8 Fan-Problem too. It would be nice if the loader could fix this. I can say that many germans do dsm on a microserver gen8
Have you read this?: http://www.silentpcreview.com/article1377-page8.html
Yes. To reduce noise you have to use raid mode. To use Raid mode you must have the driver in Xpenology - hence the requst to include it
-
I have been looking into the noice/fan issue of the Microserver G8 (Yes - it still exists with the latest bios 15% vs 6%).
One thing that i think could be interesting to test is to include the Binary only driver for the Raid Controller (https://launchpad.net/~hp-iss-team/+arc ... hp-storage).
We know that the fan in the G8 is controlled through ILO and that the ILO needs this driver to be able to get drive temp data.
Since the Controller do support creating four Raid0 arrays of one disk each we still could let the OS do the true raid work.
Just my 2c
-
as stated it will be the network that will be the true bottleneck.
For not to random read patterns 2-3 7200 rpm drives will fill a 1GB link.
If you storage needs aren't that extreme I would say 3 cheap 4 port sata cards (around $35) plus the ports from the mainboard and you got support for 16-20 drives with most mainboards
-
I just did a testinstallation on Hyper-V 2012R2. It was very easy
Create a gen 1 virtual machine using the defaults.
Download the 5592 boot.iso and mount it in your VM.
Boot the VM and do the rest from Synology Assistant just like you would on a physical server.
Install the base os first, then update 2 and 4.
Do an extra reboot after update 4 and you will be good to go
-
And I changed the graphics adapter to an old radeon 7100.
Now it boots with NoModeEmu
-
I should have an ATI card somewhere, Easy to test
-
Mine hangs at the NFS resolver
-
Mine boots with NoModeEmu and i could uppgrade it to 5.2 but after that i get a completly white webpage and thats it...
Edit: Just Wiped it and did a fresh install. Now runing DSM 5.2-5592 Update 4
Thanks Trantor
Did yours boot to the DiskStation login prompt (it went past booting to kernel)?
Thanks.
After I aplied the NoModeEmu - Yes. It gets to the login prompt without any problem.
Before the NoModeEMu it hanged at loading /zImage ...
I should add that this is on a zotac board. Haven't tested the Gigabyte one but I can do that rather Quickly.
Edit: The Gigabyte board doesn't boot with NoModeEmu. Gets to booting Kernel and hangs
-
Mine boots with NoModeEmu and i could uppgrade it to 5.2 but after that i get a completly white webpage and thats it...
Edit: Just Wiped it and did a fresh install. Now runing DSM 5.2-5592 Update 4
Thanks Trantor
Did yours boot to the DiskStation login prompt (it went past booting to kernel)?
Thanks.
After I aplied the NoModeEmu - Yes. It gets to the login prompt without any problem.
Before the NoModeEMu it hanged at loading /zImage ...
I should add that this is on a zotac board. Haven't tested the Gigabyte one but I can do that rather Quickly.
-
Mine boots with NoModeEmu and i could uppgrade it to 5.2 but after that i get a completly white webpage and thats it...
Edit: Just Wiped it and did a fresh install. Now runing DSM 5.2-5592 Update 4
Thanks Trantor
The New Active Backup
in Synology Packages
Posted
Worked nicely. Thank you