Hi !
I've been trying to switch from a mellanox single SFP+ 10GB card to an HP dual SFP+ 10GB card based on the QLOGIC 8214 chipset.
It appears the 8200 series chipset should be supported from the qlxgb driver in FreeBSD list that's linked from FreeNAS requirements.
But ! When plugged into the server, it goes like this :
- 1 if not setup and not plugged to anything, the server boots up fine
- 2 if not setup and plugged into the network, the server boots up fine, then same as 4 when trying to use the NIC
- 3 if setup and plugged, the server will throw errors everywhere until kernel panicking and shutting down when booting up
errors go like this :
qla_dump_buf8 : qla_hw_send: wrong pkt dump end
qla_hw_send: (nsegs[1, 42, 0x0] > Q8_TX_MAXSEGMENTS)
qla_dump_buf8: qla_hw_send: wrong pkt 0x2a dump start
0x00000000: b4 2e 99 3d c3 e2 a0 [...]
0x00000010 : [...]
0x00000020 : [...]
[repeat until kernel panic]
- 4 if setup and unplugged, the server will boot up fine, then, if i plug into the network, ifconfig will report the port is "inactive"
doing ifconfig down then ifconfig up on the NIC will make it work, until it stops some time later (can be 10s of minutes, or some hours, depending on load)
a service netif stop and start will generate errors like "QLA_GET_MAX_RULES command failed" and other ones like that. only a restart with unplugging and doing the ifconfig down then up will make it work again
I went back to the single port Mellanox card until this is fixed.
The hardware specs of my system should be in my signature.
The exact model of the card i'm having trouble with is the HP NC523SFP, 10GB 2P Server Adapter, Product No 593715 -- 001, RMN HSTNS - 8N64
i'm currently running 11.2-U7 up to date with the 11.2 STABLE - 11.2 [release] train.
I've been trying to switch from a mellanox single SFP+ 10GB card to an HP dual SFP+ 10GB card based on the QLOGIC 8214 chipset.
It appears the 8200 series chipset should be supported from the qlxgb driver in FreeBSD list that's linked from FreeNAS requirements.
But ! When plugged into the server, it goes like this :
- 1 if not setup and not plugged to anything, the server boots up fine
- 2 if not setup and plugged into the network, the server boots up fine, then same as 4 when trying to use the NIC
- 3 if setup and plugged, the server will throw errors everywhere until kernel panicking and shutting down when booting up
errors go like this :
qla_dump_buf8 : qla_hw_send: wrong pkt dump end
qla_hw_send: (nsegs[1, 42, 0x0] > Q8_TX_MAXSEGMENTS)
qla_dump_buf8: qla_hw_send: wrong pkt 0x2a dump start
0x00000000: b4 2e 99 3d c3 e2 a0 [...]
0x00000010 : [...]
0x00000020 : [...]
[repeat until kernel panic]
- 4 if setup and unplugged, the server will boot up fine, then, if i plug into the network, ifconfig will report the port is "inactive"
doing ifconfig down then ifconfig up on the NIC will make it work, until it stops some time later (can be 10s of minutes, or some hours, depending on load)
a service netif stop and start will generate errors like "QLA_GET_MAX_RULES command failed" and other ones like that. only a restart with unplugging and doing the ifconfig down then up will make it work again
I went back to the single port Mellanox card until this is fixed.
The hardware specs of my system should be in my signature.
The exact model of the card i'm having trouble with is the HP NC523SFP, 10GB 2P Server Adapter, Product No 593715 -- 001, RMN HSTNS - 8N64
i'm currently running 11.2-U7 up to date with the 11.2 STABLE - 11.2 [release] train.
Last edited:
