Boot sequence stalls at "no core dumps found"

[chucktryon]

I have two development systems running FreeNAS 8.3.0 (RELEASE). Both are joined to a Windows AD server (W2K8R2). Both have a problem where they sit for a *VERY* long time (15 to 20 minutes) right after the "no core dumps found". Eventually, something times out, and the rest of the system seems to boot, until it gets to "Starting crond.", where it hangs up again.... for a very long time.... and then eventually proceeds normally.

There are no reported errors on the console.

The only hints from the /var/log/messages (ie., something after a 15 minute break in the timestamps) is:

Nov 16 08:23:39 us-freenas3 kernel: bce0: Gigabit link up!
Nov 16 08:23:39 us-freenas3 kernel: bce1: Gigabit link up!
Nov 16 08:29:03 us-freenas3 kernel: mfi0: 9862 (406398544s/0x0008/info) - Battery charge complete
Nov 16 08:38:27 us-freenas3 root: /etc/rc: WARNING: failed precmd routine for vmware_guestd
Nov 16 08:38:27 us-freenas3 ntpd[2932]: ntpd 4.2.4p5-a (1)
Nov 16 08:38:48 us-freenas3 ntpd[2933]: time reset +12.518893 s
Nov 16 08:39:11 us-freenas3 smartd[3752]: Configuration file /usr/local/etc/smartd.conf parsed but has no entries (like /dev/hda)
Nov 16 08:39:11 us-freenas3 root: /etc/rc: WARNING: failed to start smartd
Nov 16 08:41:01 us-freenas3 avahi-daemon[3945]: WARNING: No NSS support for mDNS detected, consider installing nss-mdns!
Nov 16 08:54:24 us-freenas3 ntpd[2933]: kernel time sync status change 2001
​

(Incidentally, both of these systems are suffering from the "Welcome to nginx!" bug, though I've been using the command line workaround to restart the http UI.)

 

[theskyfactory]

Hello,

I am getting exactly the same error message. The CIF shares stopped working as I couldn't see the shares from windows anymore. I rebooted the machine and had to wait for a long time, which made me power it down. I powered it back on and stalled at the "no core dumps found," as you describe. It gets to the console setup but no http service is available. Did you find a fix or workaround? I would appreciate any help. Thanks.

 

[LenVolkov]

I also have run into the issue where as soon as Feenas (9.2.1.5 or even 9.2.2 - which is still beta at this time) will be added to AD and one of the folder shared (CIFS) after reboot it will be stocked at "no core dumps found" . I have also tried this but no success:
Boot in single user mode
- run this command to make the boot drive writable: mount –uw /
- run this command to edit the file: ee /etc/rc.conf
- change the line: watchdogd_enable=”YES” to watchdogd_enable=”NO”
- run this command to make the boot drive read only again: mount –ur /
- restart the PC with this command: shutdown –r now

IF anyone has any solution to that problem please advise

 

[chucktryon]

It's been a while since I've looked at this (my NAS boxes tend to go for months without rebooting!), but I believe that, at least in my case, the problem is related to having multiple, geographically isolated sites in our domain. The server seems to be trying to connect to ALL the various domain controllers in the list, including ones on networks outside of our LAN (outside of our net mask), and sitting there for at least ten seconds each waiting for the timeout. It eventually makes it through, but only after multiple passes through the list, and several minutes of not apparently doing anything.

The way around this seems to be to go to "Directory Services -> Active Directory" and either specify the exact domain controller you are going to (rather than letting FreeNAS figure out out from the SRV records in DNS), or to lower the timeout from 10s to 1s. This way, it goes directly to the correct DC rather than wasting time waiting and waiting and waiting for servers it can't reach...

(The one problem I see with this is that I have two redundant controllers to back each other up, and I don't know if there is a way to tell the FreeNAS to fail over to the second controller if the first one doesn't respond, for example, if I just rebooted it for Microsoft patch day...)

 

[LenVolkov]

Thank you SO MUC chucktryon! just one more question what version you are running now? and if you use iscsi on the same box (do you use ufs or zfs, i saw on other forum that after 9.2.2 UFS will be no longer supported http://forums.freenas.org/index.php...riate-than-zfs-on-non-ecc-motherboards.19813/ ) Thank you!

 

[chucktryon]

I have a couple of different instances of FreeNAS, each for different purposes, but this particular one is still at 9.1.1. I'm working on setting it up as our main file server for an AD based domain here. Hasn't gone into production yet, but testing looks really good!

I've been using ZFS for a long time. I probably should have more RAM on that box to really use it correctly, but it's been doing everything I need for now!

 

[LenVolkov]

Finally figured it out! Tks to chucktryon for support!
My setup HP ProLiant N40L MicroServer which has 4drives plus 2 (you can use esata to sata to add extra drive) - this hardware is 110% supported in my setup. Boot from usb
FreeNAS 9.2.1.5 x64
This time i've installed from ISO (to usb drive)
*** Make sure you follow this order:
1. add to AD
2. Use ZFS and create your folders (share them as CIFS and NFS) Ref: http://medicblog.net/?p=280
3. Reboot and make sure you can boot up
4. Create ZVOL and setup iSCSI service (if you manually choose LUN ID make sure one of you lun starts from zero, or else iSCSI service will not start/restart)
Ref: http://pawellakomski.pl/blog/definite-guide-to-iscsi-with-freenas-on-esxi-part-2/
(if you want to add additional lun then the iSCSI daemon needs to be reloaded)
now everything is working just fine!

@Chucktryon, have you figured out on how to logging to Freenas GUI with AD credentials?

 

[chucktryon]

Len, do you mean logging in to the Web GUI using AD credentials? That I haven't actually tried, since I've always used the "root" account (which was actually recently changed to a special "admin" account to separate it from root). I would imagine that it is possible to define an administrative user or group, but I haven't done the research to find out HOW to do this.