[Samba] samba_spnupdate invoked oom-killer

[steve]

On Wed, 2014-09-03 at 02:32 +0800, Chan Min Wai wrote:
> Dear Steve,
> 
> 
> Oh I See what you meant.
> Sorry. I was trying to use DRBD + OCFS2 for the sysvol which access by
> DC1 and DC2.
> After I've done that, I only found the issue about Ctdb and multi
> files locking and etc which is a bad bad idea...
> ---> Really Bad idea...
> 
> 
> Where I'm the only one using the tools and knowing that I'll be the
> one who write it.
> But still I'll kept a backup of my GPO always ;) just in case it
> crashed...
> 
> 
> But since it is working fine...
> I was thinking of why not and see how long before i hit an issue :)
> (under a single admin environment and never access and edit gpo from
> both DC)
> 
> 
> I believes that the ocfs2 and drbd are not the caused, as what i
> remember after reboot usually the drbd and ocfs2 will require a
> resync...
> 
> 
> 
> 
> 
> 
> Below are some of my ocfs2 configuration.
> 
> 
> It was similar to 
> http://wiki.gentoo.org/wiki/DRBD_with_OCFS2
> 
> 
> 
> 
> 
> /etc/ocfs2/cluster.conf
> cluster:
>         node_count = 2
>         name = amtbsysvol
> 
> 
> node:
>         number = 1
>         cluster = amtbsysvol
>         ip_port = 7777
>         ip_address = 192.168.11.23
>         name = amtbsrv01
> 
> 
> node:
>         number = 2
>         cluster = amtbsysvol
>         ip_port = 7777
>         ip_address = 192.168.11.27
>         name = amtbsrv02
> 
> 
Hi
No, I'm afraid not. We've tried it. As soon as the other node becomes
primary, it crashes spectacularly. I think they'll throw ctdb at you and
say that it doesn't work with the DC, but of course all you want to do
is sync a partition with sysvol on it. We too would like to know why
this does not work. ocfs2 supports everything ext4 supports afterall.
Given we tested this around 4.1.6, there may be reason to try again. Do
please let us know e.g. can you get both node primary and mounted at the
same time?

> 
> 
> /etc/samba/smb.conf
> # Global parameters

[   54.763131] block drbd1: role( S
> [global]
>         workgroup = WORKGROUP
>         realm = domain.net
>         netbios name = SERVER
>         server role = active directory domain controller
>         server services = s3fs, rpc, nbt, wrepl, ldap, cldap, kdc,
> drepl, winbind, ntp_signd, kcc, dnsupdate
>         idmap_ldb:use rfc2307 = yes
>         host msdfs = yes
> 
> 
>         # DNS Update
>         allow dns updates = nonsecure
> 
> 
>         #Replicate KCC Connection Object Fis
>         kccsrv:samba_kcc = false
> 
> 
>         # Log Level
>         LOG Level = 2
>         max log size = 2000
> 
> 
>         # Disable Load printer
>         load printers = No
> 
> 
>         #SMB2 Enable
>         client max protocol = SMB2
> 
> 
> [netlogon]
>         path = /var/lib/samba/sysvol/domain.net/scripts
>         read only = No
> 
> 
> [sysvol]
>         path = /var/lib/samba/sysvol
>         read only = No
> [dfs]
>         path = /home/dfsroot
>         msdfs root = yes
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> On Wed, Sep 3, 2014 at 2:18 AM, steve <steve at steve-ss.com> wrote:
>         On Wed, 2014-09-03 at 01:41 +0800, Chan Min Wai wrote:
>         > Dear Steve,
>         >
>         >
>         > What do you meant by AD Partitions?
>         > I don't have a cluster... (My 4.1.11 member files server
>         ctdb fail to
>         > run)
>         
>         Eh? You have made a drbd device primary so that you can mount
>         a file
>         system. You choose ocfs2 and you don't have a cluster? You
>         choose drbd
>         because you don't have a cluster? You mount the file system
>         successfully
>         on this node and the spnupdate crashes:
>         
>         [   54.763131] block drbd1: role( Secondary -> Primary )
>         [   61.778972] o2net: Accepted connection from node amtbsrv02
>         (num 2) at
>         192.168.11.27:7777
>         [   65.603759] o2dlm: Joining domain
>         C4BAA68B35B34A798D46CF46E2B94191
>         ( 1 2
>         ) 2 nodes
>         [   65.762744] ocfs2: Mounting device (147,1) on (node 1, slot
>         0) with
>         ordered data mode.
>         >
>         >
>         
>         > I just have 2 AD DC...
>         > This is DC1
>         > DC2 on the other hands I don't see this issue...
>         
>         
>         There is no need to cluster the dcs. You already have the
>         partitions
>         clustered as it is. Are you trying to cluster sysvol perhaps?
>         
>         Please post your ocfs2 config and that may tell us what's
>         happening a
>         litte more.
>         HTH,
>         Steve
>         
>         
>         >
>         >
>         > Thank You.
>         >
>         >
>         >
>         >
>         > On Wed, Sep 3, 2014 at 1:37 AM, steve <steve at steve-ss.com>
>         wrote:
>         >         On Wed, 2014-09-03 at 01:06 +0800, Chan Min Wai
>         wrote:
>         >         > Hello All,
>         >         >
>         >         > Anyone have seen this before?
>         >         > Did samba_spnupdate really caused the crash???
>         >
>         >
>         >         You have the AD partitions on the cluster?
>         >
>         >         --
>         >         To unsubscribe from this list go to the following
>         URL and read
>         >         the
>         >         instructions:
>         https://lists.samba.org/mailman/options/samba
>         >
>         >
>         
>         
>         
> 
>