This post is more than 5 years old
6 Posts
0
2793
Does anyone know the meaning of "isi_stats_d: connect() to devid 0 : Connection reset by peer"?
I often see the following message in /var/log/messages of some nodes.
>isi_stats_d[1893]: connect() to devid 0 192.168.0.7: Connection reset by peer
I know "192.168.0.7" is corresponding to the internal failover IP of node 7.
>nfs06-data-kks-7: lo0: flags=8049 metric 0 mtu 16384
>nfs06-data-kks-7: inet6 ::1 prefixlen 128 zone 1
>nfs06-data-kks-7: inet6 fe80::1%lo0 prefixlen 64 scopeid 0x5 zone 1
>nfs06-data-kks-7: inet 127.0.0.1 netmask 0xff000000 zone 1
>nfs06-data-kks-7: inet 192.168.0.7 netmask 0xffffff00 zone 1
Does it indicate any failure of HW or S/W?
If anyone knows the meaning of such message, please kindly tell me.
The cluster is IQ108NL x 13 nodes cluster, running with OneFS 7.1.1.11.
jepstein2
9 Posts
0
October 23rd, 2017 10:00
This is part of the routine in which isi_stats_d connects to a remote socket. isi_stats_d connects to all other nodes from all other nodes for the purpose of keeping cluster statistics updated. isi_stats_d makes these connections by way of the node name, which resolves to the node failover IP via /etc/hosts. The failover IP then resolves to either the int-a or int-b IP based on the entry in the routing table. If you see that error, it means the connect() call that isi_stats_d made to the remote socket returned a non-zero value. This could mean many things, but the two most likely situations are:
kata_TED
6 Posts
0
December 26th, 2017 01:00
jepstein-san,
Sorry for my late response. It's a very detailed explanation and I understand it. Thank you!