From the updated I/O sheet this is what's mentioned on supported SATA controllers....
Supported SAS/SATA Dual Controllers
ESX Server 3.5 supports selected SATA devices connected to dual SAS/SATA controllers. This support is limited to ESX Server 3.5. It is not included with ESX Server 3.0.x.
The supported dual SAS/SATA controllers are:
mptscsi_2xx (PCIE) ‐ LSI1078*
mptscsi_pcie ‐ LSI1068E (LSISAS3442E)*
mptscsi_pcix ‐ LSI1068 (SAS 5)*
aacraid_esx30 ‐ IBM serveraid 8k SAS controller*
cciss ‐ Smart Array P400/256 controller*
So what is the best PCIE card for me to put into my S5000PAL server that will use the above driver, if only someone could make a list :smileygrin:
Thanks,
Kev
Ok, here is another one that works with ESX 3.5 (update 3 if that matters)...
ASrock A780FULLDISPLAYPORT 780G
I did have to go through this messy process:
1) configure the sata drive as ACHI in the Bios
2) try to install, it will fail,
3) reboot and change the BIOS setting from ACHI to Non-Raid
4) the SATA drive will be seen as an IDE drive and allow you to install.
5) You won't need to do anything else, the drive will mount fine even with an IDE CD/DVD rom attached
The system uses an Athlon X2 5000 dual core chip, a 160GB Western Digital SATA drive and has 8 GB of ram. I also had to add an Intel GigE card because the MB has a realtek onboard.
All told the system came in at about $350 US. I've already built a VM using NAS storage and an NAS ISO Image, and am installing a Windows Server on it.
Happy Days.
Hi,
I also just bought a HP Proliant Ml115 G4 Quad Core Opteron but having difficult detecting the SATA Drive.
I installed ESX (not esxi) v3.5 update 3 and then copied the esx-sata-ml115.sh script using veam into /tmp and executed it but it failed.
I don't have /mnt/eroot. All i have under /nmt is cdrom.
What am i doing wrong. I also run the script in maintenance mode and normal mode.
TIA
Hi
I hope you have time to help with a litle troubleshooting ?
I have recently purchased 2 HP ML115G5s with 8GB Ram, they are superb little boxes and happily run ESXi from a USB key, I have however discovered that the disk performance is pretty poor, my assumption is that this may be to do with driver support etc ?
I have been looking for a way to get the 'full fat' version of ESX working, although as you know it installs we cannot however see the internal SATA controller.
It is at this point that I discovered a few scripts that should work but were designed for Update 1 and 2, I would love to be able to get Update3 working and as such have been trying ot get your script to work ...
I have moved the script to the local hardrive, changed permissions and executed, the script runs fine until the following section
echo "subtr lines: $lines"
I get something on teh screen that reads subtr 56 and then it just hangs, any thoughts on what might be going wrong ?
I am using the latest Update 3 Build and running the script whilst logged physically into the Service Console
Thanks
#/bin/bash
#
#
# #
HD="hda" BOOT=1 ROOT=5
INITRD_IMAGE="initrd-2.4.21-57.ELvmnix.img" COPY_INITRD_IMAGE="initrd-2.4.21-57.ELvmnix.img.gz" echo "creating directories..." mkdir -p /mnt/eboot mkdir -p /mnt/eroot echo "mounting.." mount /dev/$HD$BOOT /mnt/eboot mount /dev/$HD$ROOT /mnt/eroot mount \ | grep $HD echo "copying original initrd file..." mkdir -p /mnt/eroot/tmp/initrd/extracted cp /mnt/eboot/$INITRD_IMAGE /mnt/eroot/tmp/initrd/$COPY_INITRD_IMAGE cd /mnt/eroot/tmp/initrd echo "unpacking & mounting original initrd file..." gunzip -f $COPY_INITRD_IMAGE mount $INITRD_IMAGE extracted -t ext2 -o loop cd /mnt/eroot/tmp/initrd/extracted/etc/vmware echo "changing simple.map..." echo "10de:037f 0000:0000 sata_nv.o" >> simple.map cd /mnt/eroot/tmp/initrd/ umount extracted echo "replacing original initrd file..." gzip --best $INITRD_IMAGE mv /mnt/eroot/tmp/initrd/$COPY_INITRD_IMAGE /mnt/eboot/
/bin/cp /mnt/eboot/$INITRD_IMAGE /mnt/eboot/BACKUP-$INITRD_IMAGE /bin/cp /mnt/eboot/$COPY_INITRD_IMAGE /mnt/eboot/INITRD_IMAGE echo "patching xml files" cd /mnt/eroot/etc/vmware/pciid lines=`wc -l sata_nv.xml \ | awk ' { print $1 } '` echo "lines: $lines" (( lines -= 2)) echo "subtr lines: $lines" mv sata_nv.xml sata_nv.xml.org head -$lines > sata_nv.xml echo " <device id=\"037f\">" >> sata_nv.xml echo " <vmware label=\"scsi\">" >> sata_nv.xml echo " <driver>sata_nv</driver>" >> sata_nv.xml echo " </vmware>" >> sata_nv.xml echo " <name>MCP55 SATA Controller</name>" >> sata_nv.xml echo " </device>" >> sata_nv.xml echo " </vendor>" >> sata_nv.xml echo "</pcitable>" >> sata_nv.xml touch /mnt/eroot/etc/init.d/rp-esxcfg-pciid-boot echo "#!/bin/bash" >> /mnt/eroot/etc/init.d/rp-esxcfg-pciid-boot echo "esxcfg-pciid" >> /mnt/eroot/etc/init.d/rp-esxcfg-pciid-boot echo "sleep 5" >> /mnt/eroot/etc/init.d/rp-esxcfg-pciid-boot echo "rm /etc/rc3.d/S99rp-esxcfg-pciid-boot" >> /mnt/eroot/etc/init.d/rp-esxcfg-pciid-boot chmod a+x /mnt/eroot/etc/init.d/rp-esxcfg-pciid-boot cd /mnt/eroot/etc/rc3.d/ ln -s ../init.d/rp-esxcfg-pciid-boot S99rp-esxcfg-pciid-boot echo "cleanup..." cd / rm -f -r /mnt/eroot/tmp/initrd umount /mnt/eboot umount /mnt/eroot rm -f -r /mnt/eboot rm -f -r /mnt/eroot |
BoB
He didn't, there are a few mistakes (bugs), I now have it fixed and working, give me 30 minutes and i'll post the updated script.
This may be pointless though, I now have the controller detected and can see my storage but I get a major offline error basically as soon as you start accessing the disk it goes offline and ESX hangs
#/bin/bash #
#
#
# ###################### #
HD="hda" BOOT=1 ROOT=5
INITRD_IMAGE="initrd-2.4.21-57.ELvmnix.img" COPY_INITRD_IMAGE="initrd-2.4.21-57.ELvmnix.img.gz" echo "creating directories..." mkdir -p /mnt/eboot mkdir -p /mnt/eroot echo "mounting.." mount /dev/$HD$BOOT /mnt/eboot mount /dev/$HD$ROOT /mnt/eroot mount \ | grep $HD echo "copying original initrd file..." mkdir -p /mnt/eroot/tmp/initrd/extracted cp /mnt/eboot/$INITRD_IMAGE /mnt/eroot/tmp/initrd/$COPY_INITRD_IMAGE cd /mnt/eroot/tmp/initrd echo "unpacking & mounting original initrd file..." gunzip -f $COPY_INITRD_IMAGE mount $INITRD_IMAGE extracted -t ext2 -o loop cd /mnt/eroot/tmp/initrd/extracted/etc/vmware echo "changing simple.map..." echo "10de:037f 0000:0000 sata_nv.o" >> simple.map cd /mnt/eroot/tmp/initrd/ umount extracted echo "replacing original initrd file..." gzip --best $INITRD_IMAGE mv /mnt/eroot/tmp/initrd/$COPY_INITRD_IMAGE /mnt/eboot/
/bin/cp /mnt/eboot/$INITRD_IMAGE /mnt/eboot/BACKUP-$INITRD_IMAGE /bin/cp /mnt/eboot/$COPY_INITRD_IMAGE /mnt/eboot/$INITRD_IMAGE echo "patching xml files" cd /mnt/eroot/etc/vmware/pciid lines=`wc -l sata_nv.xml \ | awk ' { print $1 } '` echo "lines: $lines" (( lines -= 2)) echo "subtr lines: $lines" mv sata_nv.xml sata_nv.xml.old head -$lines sata_nv.xml.old > sata_nv.xml echo " <device id=\"037f\">" >> sata_nv.xml echo " <vmware label=\"scsi\">" >> sata_nv.xml echo " <driver>sata_nv</driver>" >> sata_nv.xml echo " </vmware>" >> sata_nv.xml echo " <name>MCP55 SATA Controller</name>" >> sata_nv.xml echo " </device>" >> sata_nv.xml echo " </vendor>" >> sata_nv.xml echo "</pcitable>" >> sata_nv.xml touch /mnt/eroot/etc/init.d/rp-esxcfg-pciid-boot echo '#!/bin/bash' >> /mnt/eroot/etc/init.d/rp-esxcfg-pciid-boot echo "esxcfg-pciid" >> /mnt/eroot/etc/init.d/rp-esxcfg-pciid-boot echo "sleep 5" >> /mnt/eroot/etc/init.d/rp-esxcfg-pciid-boot echo "rm /etc/rc3.d/S99rp-esxcfg-pciid-boot" >> /mnt/eroot/etc/init.d/rp-esxcfg-pciid-boot chmod a+x /mnt/eroot/etc/init.d/rp-esxcfg-pciid-boot cd /mnt/eroot/etc/rc3.d/ ln -s ../init.d/rp-esxcfg-pciid-boot S99rp-esxcfg-pciid-boot echo "cleanup..." cd / rm -f -r /mnt/eroot/tmp/initrd umount /mnt/eboot umount /mnt/eroot rm -f -r /mnt/eboot rm -f -r /mnt/eroot echo "Please Reboot Twice..." |
What do you think could be the issue with the disk going offline?
I wish I knew, I tried a brand new disk with a new VMFS, this lasted about 20 minutes before falling offline, my guess is that we may not have the right driver ?
I would be very interested to hear from anyone who has ESX Update 3 running on their ML115 G5 from local disk (non-raid) ...
I've been searching long and hard and found no one. People are mainly using ESXi on USB stick, but then they are suffering from poor disk IO. I also really want to get this to work as the server is just sitting there.
Do you think it's worth just installing version ESX v 3.5. Not 3.5 u1, u2, or u3?
I have a ML115 G1. I only run 3.5. I tried installing 3.5 u2 but I get the offline too.
I think they have changed the driver or something in the build. The reason why I think this is :-
- In 3.5 build when installing it creates VMFS on the SATA disk. When it reboots after install you get "Mount Root Failed"
- In update 2 it does not allow you to create VMFS. However it does boot up successfully after install with no errors.
Maybe in the later builds it is detecting the controller as an IDE controller. Maybe we should look for the reference of the controller in other controller files and remove them. 037f for example.
Just a thought
Looks like this may be a lot cause, I might just pull the disks and run a storage server instead, probably OpenFiler or FreeNAS, what I dont understand though is why the disk would work briefly ? longest I have managed is about 20 minutes, it could be some sort of timeout issue ? cache issue ?
Seagate drives and bad firmware?
could be, but the disks work just fine when used as a VMFS datastore with ESXi ?
it aint the disks. If it was you would get the same results with non update 3. I have even tried 2 discs one which isn't a seagate. Same with both
So we have 2 problems
ML115G with ESXi - SATA recognised for VMFS but poor performance
ML115G with ESX - SATA recognised after hack and great performance, but causes SCSI errors and eventually hangs ...
If we could fix either problem I would be a very happy bunny ...
Ich bin zur Zeit nicht im Hause. Anfragen bitte an Helpdesk Network (helpdesk@pc-ware.ch) weiterleiten!
Besten Dank
Esx 3.5 on Gigabyte EG45M-DS2H - sata ICH10 works
- in bios set Sata as AHCI
- add:
<device id="3a22">
<vmware label="scsi">
<driver>ahci</driver>
</vmware>
<name>82801JI (ICH10 Family) SATA AHCI Controller</name>
</device>
to the file /etc/vmware/pciid/ahci.xml
exectuce command esxcfg-pciid, then restart ESX server.
New disks should be accesible from ESX.
How to find device_id ???, best way is to use linux with updated pci.ids - http://pciids.sourceforge.net/
Regards,
Wieslaw
anybody successfully running ESX3.5 (not ESXi) on Shuttle SG45H7 barebone (http://global.shuttle.com/product_detail_spec.jsp?PI=1131# ) ?
as asus asus p3-p5g33 isn`t available anymore (wanted that because of http://www.thincomputing.net/articles/my-ultimate-esx-whitebox-small-aesthetically-pleasing-and-unde... i`m desparately seeking for another neat barebone which supports at least 8gb and quad cpu.
unfortunately, it`s damn hard to find one.
most barebones only have 2 slots for ram and for those with more, i have a hard time finding information if esx will run on that.
SG45H7 looks like a promising replacement, even better: you can have 16gb ram in that box.
maybe somebody has a recommendation for another small ESX box with at least 8gb ram and quad cpu ?
Hi Devzero,
I think by looking at the shuttle it will work but with a few exceptions. The Intel ICH10 controller is ESX compatible so SATA disks will work however, the NIC is a Marvell so it will need disabling in bios. Also it is likely that the IDE devices are not connected to the Intel chip either so you will need a SATA DVD drive for the install of ESX, you can leave it in there but it takes up a SATA port for an additional disk.
Lastly the RAID feature of the Intel ICH10 is FAKE raid so it will not work in ESX. It will only show as JBOD.
I hope this helps.
Cheers
Grimmy
I have now the Gigavyte EG45M-DS2H too, but the how u can install ESX(i) on it?
My esxi installation hangs, where i can choose "repair""install""ESC"
The Full ESX cant detect the Storage.
regards
meisterw