savefs: cannot retrieve client resources

Did you run that savefs? Because in my case savefs does not work that way. Actually with those arguments it fails at the end.

[root@hcrvelin]/#savefs -D9 -s hcrvelin -c hcrvelin
savefs:
Calling clnttcp_create function
savefs:
Calling clnttcp_create function
savefs: socket bound to port 14375
savefs: socket bound to port 29950
savefs: RPC Authentication: an old auth server issued the cred, it is not checksummed
savefs: lgto_auth for `nsrd' succeeded
savefs: RPC Authentication: Client successfully authenticated using AUTH_LGTO
savefs:
clu_init_lc(): ENTRY...
savefs:
get_lc_fspath_vhost_map(): ENTRY ...
savefs: No access to file: /usr/sbin/lcmap ...
savefs:
clu_init_lc(): Can't build fspath_vhost_map...
savefs:
dump_map_lc(): ENTRY ...
savefs: Lc_use_local_vhost_list = FALSE
savefs: MOUNTED filesystems
savefs:
dump_map_lc(): EXIT ...
savefs:
clu_is_cluster_host_lc(): ENTRY ...
savefs: Cannot find the Sun Cluster haget utility.savefs: Use $HA_UTIL to point to the location of the High Availability utilities.
savefs: getlocalhost_lc(): returning value: getlocalhost()
savefs: mounts[0]=/appl
savefs: lg_lstat(): Calling native lstat().
savefs: mounts[1]=/
savefs: lg_lstat(): Calling native lstat().
savefs: lg_stat(): Calling native stat().
savefs:
Calling clnttcp_create function
savefs: socket bound to port 10514
savefs: lgto_auth: redirected to hcrvelin prog 390103 vers 2
savefs:
Calling clnttcp_create function
savefs:
Calling clnttcp_create function
savefs: socket bound to port 17577
savefs: socket bound to port 10529
savefs: RPC Authentication: an old auth server issued the cred, it is not checksummed
savefs: lgto_auth for `nsrmmdbd' succeeded
savefs:
Calling clnttcp_create function
savefs:
Calling clnttcp_create function
savefs: socket bound to port 15416
savefs: socket bound to port 24100
savefs: RPC Authentication: Client successfully authenticated using AUTH_LGTO
savefs: lg_stat(): Calling native stat().
savefs: Must specify exactly one path unless using the -p option
usage: savefs [ ] path
usage: savefs -p [ ] [path ...]
options: [-BEFnqRv] [-s server] [-c client] [-l level | -C schedule] [-N name]
[-g group] [-e expiration] [-f filename] [-W width] [-t date]

However, if you try -p you may get better results:
[root@hcrvelin]/#savefs -p -s hcrvelin2 -c hcrvelin -D9
savefs: clu_is_localvirthost(): unknown cluster type
savefs: lg_fstat(): Calling native fstat().
savefs: lg_fstat(): Calling native fstat().
savefs: lg_fstat(): Calling native fstat().
savefs: lg_fstat(): Calling native fstat().
savefs:
Calling clnttcp_create function
savefs:
Calling clnttcp_create function
savefs: socket bound to port 18304
savefs: socket bound to port 20226
savefs: RPC Authentication: an old auth server issued the cred, it is not checksummed
savefs: lgto_auth for `nsrd' succeeded
savefs: RPC Authentication: Client successfully authenticated using AUTH_LGTO
savefs:
clu_init_lc(): ENTRY...
savefs:
get_lc_fspath_vhost_map(): ENTRY ...
savefs: No access to file: /usr/sbin/lcmap ...
savefs:
clu_init_lc(): Can't build fspath_vhost_map...
savefs:
dump_map_lc(): ENTRY ...
savefs: Lc_use_local_vhost_list = FALSE
savefs: MOUNTED filesystems
savefs:
dump_map_lc(): EXIT ...
savefs:
clu_is_cluster_host_lc(): ENTRY ...
savefs: Cannot find the Sun Cluster haget utility.savefs: Use $HA_UTIL to point to the location of the High Availability utilities.
savefs: getlocalhost_lc(): returning value: getlocalhost()
savefs: mounts[0]=/appl
savefs: lg_lstat(): Calling native lstat().
savefs: mounts[1]=/
savefs: lg_lstat(): Calling native lstat().
savefs: lg_stat(): Calling native stat().
savefs:
Calling clnttcp_create function
savefs: socket bound to port 17179
savefs: lgto_auth: redirected to hcrvelin2 prog 390103 vers 2
savefs: clu_is_localvirthost(): unknown cluster type
savefs: lg_fstat(): Calling native fstat().
savefs: lg_fstat(): Calling native fstat().
savefs: lg_fstat(): Calling native fstat().
savefs: lg_fstat(): Calling native fstat().
savefs:
Calling clnttcp_create function
savefs:
Calling clnttcp_create function
savefs: socket bound to port 12092
savefs: socket bound to port 14720
savefs: RPC Authentication: an old auth server issued the cred, it is not checksummed
savefs: lgto_auth for `nsrmmdbd' succeeded
savefs: clu_is_localvirthost(): unknown cluster type
savefs: lg_fstat(): Calling native fstat().
savefs: lg_fstat(): Calling native fstat().
savefs: lg_fstat(): Calling native fstat().
savefs: lg_fstat(): Calling native fstat().
savefs:
Calling clnttcp_create function
savefs:
Calling clnttcp_create function
savefs: socket bound to port 13056
savefs: socket bound to port 11190
savefs: RPC Authentication: Client successfully authenticated using AUTH_LGTO
savefs: lg_stat(): Calling native stat().
type: NSR client description;
pools supported: Yes;
browse time supported: Yes;
multiple balanced streams supported: Yes;
remote user: root;
groups: other, root, bin, sys, adm, uucp, mail, tty, lp, nuucp, daemon;
arch: sparc;
client OS type: Solaris;
CPU type: sun4u;
CPUs: 1;
kernel arch: sun4u;
machine type: workstation;
MB used: 3024;
NetWorker version: 7.2.1.Build.314;
OS: SunOS 5.9;
version: 7.2.1.Build.314;
save set: path=/, level=full, diskno=1, max_sessions=1, stype=save,\
path=/appl, level=full, diskno=0, max_sessions=1, stype=save ;
parallelism: 8
[root@hcrvelin]/#

So, here we see the file systems.

Usually you really see this kind of message when you do the typo in client name or with name resolution, but that doesn't seems to be the case here. As savefs will do the discovery (query /etc/vfstab on Solaris and /etc/fstab on the others) if you use All it seems like problem is somewhere there. I guess those file systems have their ownership and perhaps that is what is causing the issue. My understanding is you have cluster on client as well - did you run networker.cluster script (btw, if this is VCS cluster client you will need 7.2.1 on your client side)? Does the lcmap file look correct (I could see no access errors)? Are remote access set correctly on virtual/physical clients, etc... (usual cluster stuff).

With the -p option it returns similar output.

My understanding is that it has been set up for some time correctly and backing up okay; mminfo output indicates regular file system backups until 2/2/6... We are on Networker 7.1 build 230 (SCG indicates Veritas cluster is supported).

Its a very good chance that no changes have been made as they have an incredibly strict change control process here, but I won't state it categorically!

Hi David,

It is supported but:
a) being on 7.1 build 230 is asking for trouble - go to 7.1.4 if you wish to stay with 7.1.x tree (according to last update matrix 7.1 would not be supported)
b) 7.2.1 does address some issues with LCMAP under VCS - I know this as I had that last year

I just had a second client do the same thing. As soon as I replaced 'all', it's now backing up.

Hi David,

I have the same problem with a new Server 2003 client on my Networker 7.3.1 server which is running on a Server 2003 box.

It only works if I replace 'All' with it's saveset list:
SYSTEM STATE:\
SYSTEM FILES:\
SYSTEM DB:\
ASR:\
C:\
E:\
G:\
W:\

Weird!

I recently had similar issue and support asked me to upgrade to 7.2.2. Things are working good since then.

I was told by my support contractor to upgrade both the server and the clients from v7.3.1 to v7.3.2 and that did the trick. I can now use 'All' in the Savesets field.