upcp breaking named?

[arhs]

The fix worked but I am still having problems with the bind

Bind reloading on xxx using rndc zone: [xxx]
Error reloading bind on xxx: rndc: connection to remote host closed
This may indicate that the remote server is using an older version of
the command protocol, this host is not authorized to connect,
or the key is invalid.

Zone Modified!

It saves the changes but bind will not reload the changes !

I added my ISPs nameserver to the resolv.conf to fix the initial problem could this now be causiing this?

Any ideas guys?

Thx All

Try this:

pico -w /etc/rndc.key

Change it to:

key "rndc-key"

And then

pico -w /etc/named.conf


keys { "rndc-key";

Save changes and then

service named restart

 

[atsmark]

The up2date -u bind-libs command got my named running again.

I have the same problem with the bind reconfiguring error, and my rndckey has no dot or dash in it in either file, so that doesn't appear to be the issue. Help.

 

[atsmark]

Ok, I then ran /scripts/fixrndc which generated this output:

Checking rndc by querying status
Checking in /etc/named.conf to rcs system
Controls keyname does not match rndc.conf keyname
Creating /etc/rndc.key ...
Adding controls clause ...
Restarting Bind
Starting named: [OK]

And this apparently took care of the problem. I edited a DNS zone and bind reconfigured ok using rndc on both my local and remote dns servers.

Use at your own risk, it's just something I tried without knowing the details behind it.

Mark

 

[cappi]

What did it for me was:

up2date -i bind-libs

 

[nettigritty]

if you still have issues,

[[email protected]]# rpm -qa | grep bind
bind-libs-9.x.x
bind-utils-9.x.x
bind-9.x.x

make sure x.x shows the same for all these RPMs. If not, also run

up2date -u bind

 

[dreamwiz]

Was this bad update by RHE or Cpanel? Neither way, it's unacceptable to send out this kind of updates... and Im seriously considering dropping the one who is responsible. I mean is anyone testing this stuff?

 

[budway]

I don't think this was a rhe problem.


I can't understand what "stable" means any more.

 

[chirpy]

It's a combination of an RHE problem and not having disperate resolvers in resolv.conf. Nothing to do with cPanel whatsoever.

 

[janderk]

A system administrator from The Planet said in this post that CPanel was part of the problem. Apparently the cpanel sysup script updates bind, bind-devel and bind-utils but *not* bind-libs, thus breaking bind. It may be a good idea for the cpanel guys to solve this.

And yes I did learn that it is a good idea to have an additional outside name server in resolve.conf.

Jan Derk

 

[chirpy]

Fair enough, although up2date is meant to resolve dependencies and so should have automatically installed the additional module.

Either way, you should log it in bugzilla to bring it to cPanel's attention.

 

[janderk]

Either way, you should log it in bugzilla to bring it to cPanel's attention.

It seems that another The Planet sys admin already did so:

http://bugzilla.cpanel.net/show_bug.cgi?id=4402

 

[chirpy]

Yup. It also appears to bear out the fact that RH have goofed with the update.

 

[shopcentar]

Hello,

if you have servers on EV1 here is one solution, "taken" from this forum, its same as there but extra information is not bad

http://forum.ev1servers.net/showthread.php?t=62780


so, we have run this (solution from cpanel and ev1 forums).

pico /etc/resolv.conf
#add this line to the end

nameserver 207.218.245.135

#exit and save

#run this
up2date -u bind-libs

#and after that
service named restart


If its ok you can remove line nameserver 207.218.245.135 that you add in your resolve.

#restar agian bind
service named restart

And if is ok go to your WHM and run service update and NEVER have cron job to make AUTOMATIC update for your cpanel
(we have run update weekly but FIRTS check cpanel FORUMS maybe its some bug posted, if yes wait for resolution)


bye, SC

 

[driverC]

I had almost 20 hours of downtime due to this and lost several customers. Monitoring didn`t alert me as Apache, MySQL etc. was running fine. Especially as it happened right after I went to sleep. The next moring it had been down for 10 hours when I fixed it. Then one evening later it happened again and another 10 hours of downtime did occur. Now I did turn of all automated updates. Also I DID have other resolvers in the resolv.conf (I used EV1s resolvers) and the problem did still occur. Very disturbing and cost me a lot of money.

I`m also pissed to see that people who have been hosting with you for over 2 years are just running away because of such an issue after you have provided a good service to them for more than 2 years.

 

[chirpy]

I`m also pissed to see that people who have been hosting with you for over 2 years are just running away because of such an issue after you have provided a good service to them for more than 2 years.

Indeed. Clients can be horribly fickle like that. Unfortunately, nothing can be 100%, especially in computing, and problems do occur.

 

[shopcentar]

Hello,

just one more fix.

If you have error when create new account via WHM

Bind reconfiguring on ++++ using rndc
Error reconfiguring bind on ++++ : rndc: connection to remote host closed
This may indicate that the remote server is using an older version of
the command protocol, this host is not authorized to connect,
or the key is invalid.


run this:

/scripts/fixnamed
/scripts/fixndc
/scripts/restartsrv named

We have run new updates but we need run this extra commands via SSH.

Now its all ok when we create account, named restart and etc...

by, SC

 

[wipl]

I'm still getting the following errors inspite of trying out everything mentioned on here and theplanet.com forums.


Jul 23 06:23:33 rome named[16102]: zone version.bind/CH: has 0 SOA records
Jul 23 06:23:33 rome named[16102]: zone version.bind/CH: has no NS records
Jul 23 06:23:33 rome named[16102]: view.c:347: REQUIRE((&view->references)->refs > 0) failed
Jul 23 06:23:33 rome named[16102]: exiting (due to assertion failure)

Did someone faced the same errors ? If yes then can you let us know the possible fix.

Thanks in advance !

 

[mega023]

same here...

# /scripts/restartsrv named
Waiting for named to restart..............finished.


named has failed, please contact the sysadmin (result was "named is not running").
Jul 23 08:04:34 svr1 named: named shutdown failed
Jul 23 08:04:34 svr1 named[3992]: starting BIND 9.2.4 -u named
Jul 23 08:04:34 svr1 named[3992]: using 4 CPUs
Jul 23 08:04:34 svr1 named[3992]: loading configuration from '/etc/named.conf'
Jul 23 08:04:34 svr1 named[3992]: no IPv6 interfaces found
Jul 23 08:04:34 svr1 named[3992]: listening on IPv4 interface lo, 127.0.0.1#53
Jul 23 08:04:34 svr1 named[3992]: listening on IPv4 interface eth0, 67.15.202.16#53
Jul 23 08:04:34 svr1 named[3992]: listening on IPv4 interface eth0:1, 67.15.202.202#53
Jul 23 08:04:34 svr1 named[3992]: listening on IPv4 interface eth0:2, 67.15.202.114#53
Jul 23 08:04:34 svr1 named[3992]: listening on IPv4 interface eth0:3, 67.15.202.115#53
Jul 23 08:04:34 svr1 named: named startup succeeded
Jul 23 08:04:34 svr1 named[3992]: zone version.bind/CH: has 0 SOA records
Jul 23 08:04:34 svr1 named[3992]: zone version.bind/CH: has no NS records
Jul 23 08:04:34 svr1 named[3992]: view.c:347: REQUIRE((&view->references)->refs > 0) failed
Jul 23 08:04:34 svr1 named[3992]: exiting (due to assertion failure)



also when up2date -u bind

i'm getting error

error: unpacking of archive failed on file /usr/bin/dig: cpio: rename
There was a fatal RPM install error. The message was:
There was a rpm unpack error installing the package: bind-utils-9.2.4-14_EL3

 

[budway]

open a support ticket with your dc our cpanel they can help you out

 

[mega023]

I did but today is saturday and they are very slow...
and then they say me that "it's not their problem it's problem with cPanel"...
and stuff like that...

and my 100sites are down for 12 hours now
and I'm getting 100 phone calls and I still can't figure out this s___t