TritonDataCenter / smartos-live

For more information, please see http://smartos.org/ For any questions that aren't answered there, please join the SmartOS discussion list: https://smartos.topicbox.com/groups/smartos-discuss
1.57k stars 245 forks source link

Link Aggregation fails to setup vnics on newer releases (since ) #225

Open bayoda opened 11 years ago

bayoda commented 11 years ago

Tested with joyent-20130530T224720Z -> Machines are offline completely.

Last mostly working version (but as it seems not with mixed NIC drivers)
joyent_20130222T000747Z

I just can post the latests svcs -L network/physical (with joyent_20130222T000747Z) but with two machines - one is working the other not (nearly identical config - except IP adresses)

aggregates are there - as seen in dladm show-aggr

addititonal nics are NOT created (extern / storage / web / linux)

So here we go:

Config: Big Maschine: (failing to establish correctly) 

int0_aggr=00:25:90:95:94:b2,00:25:90:95:94:b3
ext0_aggr=68:5:ca:a:92:b6,68:5:ca:a:92:b7

admin_nic=int0
extern_nic=ext0
storage_nic=ext0
web_nic=ext0
linux_nic=ext0

admin_ip=192.168.210.100
admin_netmask=255.255.255.0
admin_gateway=192.168.210.100

extern0_ip=192.168.210.99
extern0_netmask=255.255.255.0
extern0_gateway=192.168.210.1

storage0_ip=192.168.210.98
storage0_netmask=255.255.255.0
storage0_gateway=192.168.210.1

web0_ip=192.168.210.97
web0_netmask=255.255.255.0
web0_gateway=192.168.210.1

linux0_ip=192.168.210.96
linux0_netmask=255.255.255.0
linxu0_gateway=192.168.210.1

headnode_default_gateway=192.168.210.1

dns_resolvers=8.8.8.8,8.8.4.4
dns_domain=XXXXXXXXX.com
# removed domain as this is a working system on the world wild web
ntp_hosts=pool.ntp.org
compute_node_ntp_hosts=192.168.210.100

+ echo 'ERROR: admin NIC not found, unable to bring up admin network.'
ERROR: admin NIC not found, unable to bring up admin network.
+ exit 96
[ Jun  6 15:02:37 Method "start" exited with status 96. ]
[ Jun  6 18:19:56 Executing start method ("/lib/svc/method/net-physical"). ]
[ Jun  6 18:19:56 Timeout override by svc.startd.  Using infinite timeout. ]
+ smf_configure_ip
+ /sbin/zonename -t
+ [ global = global -o shared = exclusive ]
+ return 0
+ LD_LIBRARY_PATH=/lib
+ export LD_LIBRARY_PATH
+ smf_netstrategy
+ smf_is_nonglobalzone
+ [ global != global ]
+ return 1
+ /sbin/netstrategy
+ set -- ufs none none
+ [ 0 -eq 0 ]
+ [ ufs = nfs -o ufs = cachefs ]
+ _INIT_NET_STRATEGY=none
+ export _INIT_NET_STRATEGY
+ smf_is_globalzone
+ [ global = global ]
+ return 0
+ /usr/sbin/dladm init-phys
+ log_if_state before
== debug start: before ==
LINK         MEDIA                STATE      SPEED  DUPLEX    DEVICE
igb0         Ethernet             unknown    0      half      igb0
igb1         Ethernet             unknown    0      half      igb1
e1000g0      Ethernet             unknown    0      half      e1000g0
e1000g1      Ethernet             unknown    0      half      e1000g1
lo0: flags=2001000849<UP,LOOPBACK,RUNNING,MULTICAST,IPv4,VIRTUAL> mtu 8232 index 1
    inet 127.0.0.1 netmask ff000000 
lo0: flags=2002000849<UP,LOOPBACK,RUNNING,MULTICAST,IPv6,VIRTUAL> mtu 8252 index 1
    inet6 ::1/128 
== debug end: before ==
+ load_sdc_sysinfo
/lib/svc/method/net-physical[188]: load_sdc_sysinfo[19]: eval[1]: SYSINFO_NIC_#linux=e1000g0: not found [No such file or directory]
[ Jun  6 18:19:57 Method "start" exited with status 127. ]
[ Jun  6 18:19:57 Executing start method ("/lib/svc/method/net-physical"). ]
[ Jun  6 18:19:57 Timeout override by svc.startd.  Using infinite timeout. ]
rmustacc commented 11 years ago

Thanks, this helps. Can you please also include the output of dladm show-phys -m?

bayoda commented 11 years ago

Here it is !

dladm show-phys -m LINK SLOT ADDRESS INUSE CLIENT igb0 primary 0:25:90:95:94:b2 yes int0-igb0 igb1 primary 0:25:90:95:94:b2 yes int0-igb1 e1000g0 primary 68:5:ca:a:92:b6 yes ext0-e1000g0 e1000g1 primary 68:5:ca:a:92:b6 yes ext0-e1000g1

bayoda commented 11 years ago

maybe I'm really wrong but - if I define an interface over aggregates in "OPENsolaris" days - i just configured it - but it get an IP adress in the zone ... (OS / OI) - so dladm create-vnic -l aggregate name ...

If i configure: 2 aggregates and serveral "nics" on top of smartos it works without an IP adress - but the documentation tells define your IP adress (tried both ways) seems that in smartos_ 20130222T000747Z - this works ... but not really sure if I'm correct

So
admin_nic=aggr0 admin_ip=192........ admin_netmask=255..... admin_gateway=192.....

extern_nic=aggr1 extern0_ip=192...... extern0_netmask=255..... extern0_gateway=192......

storage_nic=aggr1 storage0_ip=192..... storage0_netmask=255..... storage0_gateway=192.....

is this correct this way?