Bug in web-interface on 3.8.8

On my new, reinstalled 3.8.8 I get this:

This is a freshly install from the 3.8.7 ISO, which I installed and upgraded, and then went ahead and changed the IP to the static one I use for the NAS. After another reboot, I get different IP’s shown. I access the NAS via 172.26.1.8 (also SSH), but the Web interface shows 172.26.1.17.

I actually had to reinstall, because the original update left me in a state where the cli told me the system had no IP, and I couldn’t connect to the NAS from other machines.
So I’m worried that something is wrong with the networking setup somewhere.

@KarstenV Thanks for reportign this and yes it is a bit disconcerting. I haven’t managed to reproduce here but need to put more time into it. I know the network stuff has been reworked of late so the 3.8-7 and 3.8-8 systems are actually different. Here I have agreement on the actual and displayed ip so far (fresh 3.8-7 to 3.8-8 upgrade) but still using dhcp. Will try your steps once I’ve done with some other loose ends.

I have seen the “system has no ip message” and we have an open issue on this: bogus noip message in console In that case I was not infrequently seeing that message and after just a little wait and a carriage return and the message was updated to what might be expected. Some kind of timing issue. However the scope of the changes in the network config may well have changed this completely. I think we are going to have to wait until @suman can comment / reproduce.

Glad you got up and running again. Really we need some kind of “Find my Rockstor” web based thing which has been discussed but I just don’t know how this can be done just yet, I was thinking of some kind of network scan that looked for modern kernel signatures in ip returns. Any ideas welcome as this console message is only part of the story really.

I had this problem: https://github.com/rockstor/rockstor-core/issues/917

Can you show us your /etc/hosts ?

It could seem you are on to something :smile:

Contenst of hosts file:

127.0.0.1 localhost localhost.localdomain localhost4 localhost4.localdomain4
::1 localhost localhost.localdomain localhost6 localhost6.localdomain6
172.26.1.17 RockStor NAS

@roweryan Just chipping in, a fresh install of 3.8-7 with 3.8-8 upgrade looks like this:-
127.0.0.1 localhost localhost.localdomain localhost4 localhost4.localdomain4
::1 localhost localhost.localdomain localhost6 localhost6.localdomain6
192.168.1.203 rockstor
No network changes have been made.

@phillxnet I can only reproduce the problem when I change ip. I am using dhcp with static leases and changed the lease ip.

My DHCP server hands out adresses from 172.26.1.12 and up, and Rockstor was given .17 on install, and that was the ip I used to log in and reconfigure from DHCP to static .8. After this I can access Rockstor on .8 as I should, but the Web-interface shows the old adress.

I worry that the wrong input into hosts will have other effects on e.g. rockons?

I fixed /etc/hosts and everything was fine for me, but I only use one rock-on. @phillxnet and @suman are the experts!

Given that
cat /etc/nsswitch.conf | grep hosts
gives:-

#hosts:     db files nisplus nis dns
hosts:      files dns myhostname

Then with files first the hosts file is going to upset something or other if it gets out of kilter in your configs at one time or another.

It should be that if one does nslookup on a machine name it should resolve to it’s actual ip, both locally and remotely. Also proper if nslookup on the actual ip resolves to the machine name ie:-

nslookup rockstor.lan

Server:		192.168.1.1
Address:	192.168.1.1#53

Name:	rockstor.lan
Address: 192.168.1.179

and
nslookup 192.168.1.179

Server:		192.168.1.1
Address:	192.168.1.1#53

179.1.168.192.in-addr.arpa	name = rockstor.lan.

Here my dns server is 192.168.1.1 and hosts a local domain called lan, not all routers do this properly though.
So all looks good on my default dhcp Rockstor. @KarstenV I think you can safely change you’re /etc/hosts file to match the actual configured IP like @roweryan did as this does look like a bug to me.

However I’m going to have to pass this one over to @suman to issue / final say etc. The above stipulation is proper but not always absolutely necessary, but servers do like it that way is my understanding, especially with things like kerberos authentication and the like.

Hope that helps (nslookup is in bind-utils by the way).

1 Like

I have changed the hosts file to reflect the systems real ip-adress.

Having now had time to reboot the system, the Web-UI still shows the wrong Mgmt IP, even though the hosts file now/still shows the correct one.

If I click the IP, I am taken to the Appliences tab. There an appliance is shown (as it allways has been) for port 443, but pointing to the wrong IP.

Something is not right.

When I updated /etc/hosts and rebooted, the appliances page updated as well.

What happens if you run:

grep WRONGIP -r /etc/

on the rockstor device?

Nothing seems to happen.

The command executes without any errors, but nothing seems to be changed.

Did you swap WRONGIP for the wrong ip address showing in the web interface?

No :smile:

But I have now run the command:

grep 172.26.1.17 -r /etc/

And made a reboot, with no result.

I still connect to 172.26.1.8, but web-ui shows 172.26.1.17 as Mgmt IP.

Can you get a screenshot of the network interfaces screen, and a copy paste of /etc/hosts? Also the output of

ip addr
ip link

Yes.

Contents of hosts:
127.0.0.1 localhost localhost.localdomain localhost4 localhost4.localdomain4
::1 localhost localhost.localdomain localhost6 localhost6.localdomain6
172.26.1.8 RockStor NAS

ip addr:
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
valid_lft forever preferred_lft forever
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2: enp3s0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
link/ether 00:22:15:cf:46:8a brd ff:ff:ff:ff:ff:ff
inet 172.26.1.8/24 brd 172.26.1.255 scope global enp3s0
valid_lft forever preferred_lft forever
inet6 fdf9:c0d6:984a:0:222:15ff:fecf:468a/64 scope global dynamic
valid_lft 7014sec preferred_lft 1614sec
inet6 fe80::222:15ff:fecf:468a/64 scope link
valid_lft forever preferred_lft forever

ip link:

1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN mode DEFAULT
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
2: enp3s0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP mode DEFAULT qlen 1000
link/ether 00:22:15:cf:46:8a brd ff:ff:ff:ff:ff:ff

There is a difference between the netmask on my rockstor NAS and the ones on the rest of my network.

All my windows PC’s (and other equipment on my network) are set up with a netmask of 255.255.0.0, but rockstor reads 255.255.255.0.

It won’t let me change the netmask to match the other machines in my network. When I change it and click “Submit” it doesnt change.

Could this have something to do with the problems?

I don’t think that’s related, but that sounds like a bug.

I can reproduce your problem even after editing the hosts file and rebooting.

I updated https://github.com/rockstor/rockstor-core/issues/917

1 Like

I guess this is stored in the postgres db.