Prashanth Sundaram's questions

Over the past several days I have been trying to get the shadow to work with samba but haven’t been successful. Can someone check below config and let me know if I am missing something? We are using Equallogic SAN and iSCSI LUNS to mount volumes.

I can cleanly access samba shares on Windows 7 clients but just not shadow copy. I have referred the official how-to but couldn’t get it to work. I see these messages in the logs. Any help is deeply appreciated.

[2012/10/31 12:20:53.549863, 0] smbd/nttrans.c:2170(call_nt_transact_ioctl)
FSCTL_GET_SHADOW_COPY_DATA: connectpath /fs/test-01, failed.

[2012/10/31 12:21:13.887198, 0] modules/vfs_shadow_copy2.c:734(shadow_copy2_get_shadow_copy2_data)
shadow:snapdir not found for /fs/test-01 in get_shadow_copy_data

[2012/10/31 12:21:13.887265, 0] smbd/nttrans.c:2170(call_nt_transact_ioctl)
FSCTL_GET_SHADOW_COPY_DATA: connectpath /fs/test-01, failed.

== Samba pkgs ==

• samba-3.5.10-116.el6_2.x86_64
• samba-common-3.5.10-116.el6_2.x86_64
• samba-winbind-clients-3.5.10-116.el6_2.x86_64
• samba-client-3.5.10-116.el6_2.x86_64

=== df –h == First is the iSCSI LUN and 2 others are snapshots.

/dev/mapper/eql-0-fs-test01   5.0G  2.3G  2.5G  48% /fs/test-01 
/dev/mapper/eql-2-0+fs-test01 5.0G  2.3G  2.5G  48% /fs/test-01/@GMT-2012.10.26-17.32.42/fs/test-01   (SNAPSHOT-1) 
/dev/mapper/eql-d-0+fs-test01 5.0G  2.3G  2.5G  48% /fs/test-01/@GMT-2012.10.31-11.52.42/fs/test-01   (SNAPSHOT- 2)

===/etc/samba/smb.conf ===

[global]
        workgroup = DOMAIN
        server string = Samba Server Version %v
        security = ads
        realm = DOMAIN.CORP
        encrypt passwords = yes
        guest account = nobody
        map to guest = bad uid
        log file = /var/log/samba/%m.log

        domain master = no
        local master = no
        preferred master = no
        os level = 0

        load printers = no
        show add printer wizard = no
        printable = no
        printcap name = /dev/null
        disable spoolss = yes

        follow symlinks = yes
        wide links = yes
        unix extensions = no

[test]
        comment = Test Directories
        path = /fs/test-01
        vfs objects = shadow_copy2
        #shadow_copy2: sort = desc
        #shadow: localtime = yes
        #shadow: snapdir = /fs/test-01/test
        #shadow: basedir = /fs/test-01
        guest ok = yes
        writeable = yes
        map archive = no
        force create mode = 0660
        force directory mode = 2770
        inherit owner = yes
        inherit permissions = yes

All feedback is welcome. Thanks!

I am trying to deploy a custom registry entry which I exported from a test machine. It looks like below. I came across THIS similar request on another site, but I couldn't make it to work.

"TextFontSimple"=hex:3c,00,00,00,1f,00,00,f8,00,00,00,40,dc,00,00,00,00,00,00,\ 00,00,00,00,ff,00,31,43,6f,75,72,69,65,72,20,4e,65,77,00,00,00,00,00,00,00,\ 00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00

As per the other solution, my PS command below, throws error."A parameter cannot be found that matches parameter name"

Set-ItemProperty -Path "HKEY_CURRENT_USER\Software\Microsoft\Office\14.0\Common\MailSettings" -Name "TextFontSimple" -PropertyType Binary -Value ([byte[]] (0x3c,0x00,0x00,0x00,0x1f....0x00))

Any ideas?

====EDIT=====

The key & value already exists. When I use Get-ItemProperty

PSPath         : Microsoft.PowerShell.Core\Registry::HKEY_CURRENT_USER\Software\Microsoft\Office\14.0\Common\MailSettings
PSParentPath   : Microsoft.PowerShell.Core\Registry::HKEY_CURRENT_USER\Software\Microsoft\Office\14.0\Common
PSChildName    : MailSettings
PSProvider     : Microsoft.PowerShell.Core\Registry
TextFontSimple : {60, 0, 0, 0...}

I am working for a small shop where we want to provision all the hosts from one single location. Right now, we use Legacy RIS on Win server 2003 to deploy hosts via PXE boot. Though this method works, but I wanted to know if the community has a better way of handling this.

Though we don't use any Configuration management tool but my assumption (or rather plan to use one) is it will kick in during the last step of provisioning and will install packages/configs to bring it up to desired state.

Any good tools or methods that I can consider in a multi-OS environment?

I am preparing for a Systems Admin(linux) role for a big company(top 10 companies) and one of the requisite is knowledge of a language like perl, python or Ruby. I have a phone interview which involves sharing a editor with the interviewer, now what kind of coding exercise can I expect on a phone interview or a regular interview?

I have used shell for most of my day-to-day tasks and occasionally Perl and python for bigger tasks. Now I don't expect them to ask

Write a script that syncs AD users to LDAP.

I don't think data structures, algorithms or sorting questions apply to me role? Can you give me any examples of SysAdmin related coding questions?

This is every SysAdmin's nightmare of things to do. Basically we want to control who has access to which hosts. As simple it may sound, the problem is to find a scalable and low maintenance(mgmt. overhead) solution. We use bcfg2 for Config Mgmt much like Cfengine & puppet.

Some ways:

1. Netgroups is very scalable but comes with a huge mgmt overhead. Maintaining hosts, group of hosts, user netgroups(separate from ldpa groups) seems like a very big burden, but is doable.

2. ldap.conf (go to the post by jmozdzen on 1st July) and LDAp based access control. We could template the ldap.conf for each host and create a group with hostname and members as users. But the downside is you cannot specify ldap group(team of users) for access but only individually.

3. sshd_config restriction. But that doesn't work if the users log in natively.

4. Host attribute check. By un-commenting pam_check_host_attr in ldap.conf and add the hostname to each user works well, but automation is not easy.

Anybody has a different approach to this problem and that scales well and automated?

I have a snmpd.conf which looks like this. I would like the comments to auto-fill i.e. make the file human readable with explanations. If it were one file, I could have done copy/paste from default one, but there are many different ones. Is there a command that can insert the comments?

snmpconf -R [belowfile] -a -f [output file]

Above command didn't help :(

m2sec notConfigUser  default       public

group   notConfigGroup v1           notConfigUser
group   notConfigGroup v2c           notConfigUser

view    systemview    included   .1.3.6.1.2.1.1
view    systemview    included   .1.3.6.1.2.1.25.1.1

access  notConfigGroup ""      any       noauth    exact  systemview none none

com2sec local     localhost       bl4h
com2sec wgen 10.1.0.0/22      bl4h

group MyRWGroup  any        local
group MyROGroup  any        DOMAIN

view all    included  .1                               80

view mib2   included  .iso.org.dod.internet.mgmt.mib-2 fc

access MyROGroup ""      any       noauth    0      all    none   none
access MyRWGroup ""      any       noauth    0      all    all    all

access  notConfigGroup ""      any       noauth    exact  roview rwview none

syslocation SFO
syscontact sa.unix <[email protected]>

disk / 10000

load 12 14 14

pass .1.3.6.1.4.1.4413.4.1 /usr/bin/ucd5820stat

The reason I posted this question is I read somewhere that snmpconf can insert comments and make it look more understandable. I have a bunch of these files and wanted to make it look pretty and understandable by Jr. SAs.

We currently have LDAP servers in Dev env. and staging/prod env. Both are not coupled and accounts are created manually. So the uids are same but the uidNumber and gidNumber are different across these two envs.

I am trying to upgrade to 389 DS and setup replication, so that users are sync'ed from Dev to staging/prod ldap servers without needing an admin. But doing that will involve changing the uidnum/gidnum but the userid will stay same. I wrote a script which gives me the list of users with conflicting uidnumbers and it says I have around 250 users like that. Sample:

amauer STG:2596 | DEV:2581  conflictID:aling
aling STG:2581 | DEV:2678  conflictID:blsow
bslow STG:2678 | DEV:2625  conflictID:krusso

As you can see amauer with staging uidNumber=2596 has 2581 in dev. If I go ahead and change it via script below, I would mess up files that belong to amauer when i do the same on aling user. As at that time, amour and aling will have same uidNumber.

cat "$chownlist" | xargs -0 sudo chown $user
cat "$chgrplist" | xargs -0 sudo chown :$group
sudo rm "$chownlist" "$chgrplist"

Did anyone have similar problem while bulk changing file owner/group? Some kind of sorting mechanism which will come up with a list that

This project will give me 2-3hrs extra time every week and automation is OUR profession. :)

Any ideas??

Thanks