TOPIC:

Hi,

Am doing to a load of testing with the latest ha-lizard and ha-iscsi release. I've installed the NoSAN installer 2.1.1

Everything is working, but when I do a iscsi-cfg become-secondary we get an error (see attached output), then the server sits for 1 minute 30 seconds before it acknowledges that it has become secondary.

The error we get is:
[root@MYSERVER1 ~]# iscsi-cfg become-secondary
/usr/bin/iscsi-cfg: line 275: 1557 Terminated /etc/iscsi-ha/iscsi-ha.sh become_secondary

We think we are doing everything correctly; e.g.
Disable HA
Put iscsi into manual mode on BOTH servers
performing iscsi-cfg become-secondary on the PRIMARY server.

Can you tell me if something is wrong, or is this just a warning that can be ignored? I am not seeing any issues when doing this, apart from the warning message and the 1 minute 30 seconds wait.

I have attached the putty output from both servers. SERVER1 is Pool Master, SERVER2 is the Slave.

Many thanks in advance for help.

Adam

There is a new release that I plan on posting this weekend that addreses this. In the meantime, try “systemctl daemon-reload” on both hosts and check if the transition time improves when swapping roles in manual mode. The changeover should be approximately 5 seconds.

Also, this release addresses an arp issue when using a bond as the replication link. There were cases where the floating IP was not advertised to the peer after a simulated crash.

Hi, great news, I look forward to trying the new version!

Adam

Hi,

I've installed both new versions of ha-lizard (ha-lizard-2.1.2) and iscsi-ha (iscsi-ha-2.1.1).

The error message when issuing the command "iscsi-cfg become-secondary" has gone, but it still takes 30 seconds to respond, stating that the node is still primary. After 1 minute 30 seconds and reissuing the command "iscsi-cfg become-secondary" the node shows that it has successfully transitioned to become the secondary node.

Please see putty output files attached.

Regards,

Adam

Have you issued "systemctl daemon-reload" on both servers?

Salvatore Costantino wrote: Have you issued "systemctl daemon-reload" on both servers?

Sorry didnt know I needed to, I thought the versions I installed addressed this? I will give it a try now.

thanks,

Adam