FreeNAS-9.3-STABLE-201506292130
Manufacturer: Supermicro
Product Name: SSG-5028R-E1CR12L
HBA - LSI Logic / Symbios Logic SAS3008 IT Mode (imbedded)
32G RAM
12 Drives - WD Red Nasware 3.0 3TB
This server is set up with a single pool of mirrors using 12 drives and a single ssd cache. OS is also on a separate ssd drive. The pool drives will randomly encounter write errors causing it to be marked as faulty. Sometimes I will loose two drives in separate mirrors at the same time. Luckily I've not lost two drives in a single mirror yet. Performance is fine and the swap file is never used.
I'll replace it with another known good drive, then several days, sometimes up to a week will go by before another drive is marked faulted. There is no common drive slot which repeatedly has this issue, the HBA has been updated to the latest available firmware and there are no entries in the syslogs related to drive errors. Memory tests are good as well. The drives marked as faulty always pass smartctl tests. This has been going on for several months.
I've got another identical server running ZFS for Linux on Debian and have not experienced this problem so far.
Any ideas?
Thanks
Manufacturer: Supermicro
Product Name: SSG-5028R-E1CR12L
HBA - LSI Logic / Symbios Logic SAS3008 IT Mode (imbedded)
32G RAM
12 Drives - WD Red Nasware 3.0 3TB
This server is set up with a single pool of mirrors using 12 drives and a single ssd cache. OS is also on a separate ssd drive. The pool drives will randomly encounter write errors causing it to be marked as faulty. Sometimes I will loose two drives in separate mirrors at the same time. Luckily I've not lost two drives in a single mirror yet. Performance is fine and the swap file is never used.
I'll replace it with another known good drive, then several days, sometimes up to a week will go by before another drive is marked faulted. There is no common drive slot which repeatedly has this issue, the HBA has been updated to the latest available firmware and there are no entries in the syslogs related to drive errors. Memory tests are good as well. The drives marked as faulty always pass smartctl tests. This has been going on for several months.
I've got another identical server running ZFS for Linux on Debian and have not experienced this problem so far.
Any ideas?
Thanks