BlindOracle
Cadet
- Joined
- Sep 30, 2017
- Messages
- 9
Dear community,
we bought a brand new Intel server, meat to be used as ESX Datastore, running FreeNAS.
Here are the specs:
Intel Server System R2224WFTZS
2 x 10GBit LAN (On Board)
2x Intel Xeon Bronze 3104 - 1,7 GHz / 8,25M / 6 Core / 85W
2x 32GB (ECC Registered DDR4 2666Mhz)
BROADCOM LSI SAS3 9300-4i SGL 12GB/s HBA / Controller
Intel 36-Port SAS3/SATA Storage Expander (RES3TV360)
2x Toshiba Allegro14 SAS 300GB (15k)
6x Toshiba Allegro14 600GB (10k)
9x Toshiba VX500 SSD 1TB
FreeNAS 11.1-U6 is installed on a mirror on the 2 300GB drives.
The 6 600GB drives are configured as RAIDz and 8 SSDs are configured as RAID-10 pool with one spare drive.
So far so good, everything turned out to be smooth, until we started to use the SSD-Pool on ESX.
Either mouted via iSCSI or as NFS-Share, as soon as we start to migrate a VM onto that DS, the pool starts throwing disk failured, becomes degraded and finally fails completely with at least 5 SSDs. I already installed the latest IT-Firmware on the 9300-Controller, made no difference.
SMART-Tests showing no errors on the SSDs, also Toshiba tells that these drives should run fine, even they are not server-class SSDs.
When we put these disks into the old server, which is a SuperMicro brand, they do run smoothly and don't show errors when used as Datastore.
Currently we're re running out of ideas.
BTW: Using the SASPOOL as Datastore works fine, it's just too slow. :-(
Anyone over here, any ideas?
Thanks very much in advance
Blind
we bought a brand new Intel server, meat to be used as ESX Datastore, running FreeNAS.
Here are the specs:
Intel Server System R2224WFTZS
2 x 10GBit LAN (On Board)
2x Intel Xeon Bronze 3104 - 1,7 GHz / 8,25M / 6 Core / 85W
2x 32GB (ECC Registered DDR4 2666Mhz)
BROADCOM LSI SAS3 9300-4i SGL 12GB/s HBA / Controller
Intel 36-Port SAS3/SATA Storage Expander (RES3TV360)
2x Toshiba Allegro14 SAS 300GB (15k)
6x Toshiba Allegro14 600GB (10k)
9x Toshiba VX500 SSD 1TB
FreeNAS 11.1-U6 is installed on a mirror on the 2 300GB drives.
The 6 600GB drives are configured as RAIDz and 8 SSDs are configured as RAID-10 pool with one spare drive.
So far so good, everything turned out to be smooth, until we started to use the SSD-Pool on ESX.
Either mouted via iSCSI or as NFS-Share, as soon as we start to migrate a VM onto that DS, the pool starts throwing disk failured, becomes degraded and finally fails completely with at least 5 SSDs. I already installed the latest IT-Firmware on the 9300-Controller, made no difference.
SMART-Tests showing no errors on the SSDs, also Toshiba tells that these drives should run fine, even they are not server-class SSDs.
When we put these disks into the old server, which is a SuperMicro brand, they do run smoothly and don't show errors when used as Datastore.
Currently we're re running out of ideas.
BTW: Using the SASPOOL as Datastore works fine, it's just too slow. :-(
Anyone over here, any ideas?
Thanks very much in advance
Blind
