Tomasz Chmielewski schrieb:
> Niels de Carpentier schrieb:
>> I have a number of systems with an iscsi root filesystem. These systems
>> connect to an redundant pair of iscsi servers, using tgtd. I use heartbeat
>                                                         ^^^^
>> Is there any way to prevent this, so a iscsi root system can recover
>> gracefully from a short outage?
> You are using tgtd.
> What are the causes of these disconnections? I guess manual caused by 
> you, but is it:
> 1) cabling, switches
> 2) firewall, routing etc.
> 3) tgtd restart?

Anyway, if it's 3) tgtd restart - it's "by design", and you should 
complain on stgt-devel mailing list. There were some changes in the git 
lately, but AFAIK, it hasn't improved in all areas.

If it's either 1) or 2), there is something fishy here.

Supposing it's 3) - there is a slight race before you start tgtd and 
before you can configure targets with tgtadm. Therefore, a "sleep 2s" 
would be recommended. The rest is like below - tgtd already listens on 
3260, but has no targets configured. Any initiator that connects will be 
rejected, and hence your immediate fs errors.

One workaround is to block iSCSI traffic with iptables before starting 
tgtd and removing the block after all targets are configured.

target              initiator
not started         trying to reconnect
start tgtd          trying to reconnect
sleep 2s            trying to reconnect
nothing configured  login I/O error - non fatal
configure target1   conn to target1 OK
no such target      conn to target2 FAIL
                      I/O error to target2
configure target2   too late, fatal, we lost it

Tomasz Chmielewski

You received this message because you are subscribed to the Google Groups 
"open-iscsi" group.
To post to this group, send email to
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at

Reply via email to