Backup Central

Legato Admins

My Unix directives include this line:

<< / >>
+skip: core nohup.out crash*

But, core files are getting backed up. This seems wrong to me. Does
anyone have any idea why?

Environment:
HP-UX 11.00 NetWorker Server
NetWorker 7.1.2

Thanks,
Pat

Note: To sign off this list, send a "signoff networker" command via email
to listserv < at > listserv.temple.edu or visit the list's Web site at
http://listserv.temple.edu/archives/networker.html where you can
should be sent to stan < at > temple.edu

When you say they're getting backed up, which ones are you referring to?
The ones under the root file system (/) itself or some other file
system? I would think the "+skip" will only skip core files located
anywhere under / but not say /disk2, /home, etc. if these were in fact
separate file systems. Is that the case?

Also, how long is your directive? How many lines do you have? Is this a
server side directive or a client side using .nsr file? I know there
used to be a limitation in how long a server side directive could be. I
don't recall what the limit in total number of characters was, but may
no longer be a problem with more recent releases?

Could also try removing the directive and re-keying. Could be some
hidden jibberish is in there causing mischief. I always use tabs
preceeding the 'skip', but spaces should work. I use spaces on .nsr
files, and they work there.

Also, is there by any chance a .nsr file under / that's overriding this
by including the 'always' uasm or something like that?

George



Pat DeCelie wrote:

Legato Admins

My Unix directives include this line:

<< / >>
+skip: core nohup.out crash*

But, core files are getting backed up. This seems wrong to me. Does
anyone have any idea why?

Environment:
HP-UX 11.00 NetWorker Server
NetWorker 7.1.2

Thanks,
Pat

--
Note: To sign off this list, send a "signoff networker" command via email
to listserv < at > listserv.temple.edu or visit the list's Web site at
http://listserv.temple.edu/archives/networker.html where you can
should be sent to stan < at > temple.edu

Note: To sign off this list, send a "signoff networker" command via email
to listserv < at > listserv.temple.edu or visit the list's Web site at
http://listserv.temple.edu/archives/networker.html where you can
should be sent to stan < at > temple.edu

George

Yes, the core files were backed up on a filesystem that is not the root FS.
However, this means I have to put the "+skip: core" for each filesystem.
That is a bummer. As for the length, I think it was 1024 chars until rev.
7 of NetWorker. So, that doesn't apply here. I'll check for jibberish.

Thanks,
Pat

Note: To sign off this list, send a "signoff networker" command via email
to listserv < at > listserv.temple.edu or visit the list's Web site at
http://listserv.temple.edu/archives/networker.html where you can
should be sent to stan < at > temple.edu

Well, I'm not categorically stating that you have to. That was just my
surmise as to how it's little brain works. I mean, for example, if I
list say / under the savesets for a given client then is NetWorker gonna
back up just / or is it gonna back up all the file systems? Of course,
we all know it will descend no further than /; so, with the exception of
the mount point path names, any other file systems would not be captured
even though their mount points are under the root FS. I guess I was
thinking that directives would follow along the same logic?

But I would try it, and if doing it that way works ... natch. Should be
easy enough to test. As far as it being a bummer ... yes, it is. Might
be something in the man page, though, for uasm that might tell you how
to work around it?

Might be a nice enhancement for future releases. Say, how about if you
use << All >>. Does NetWorker recognize that? Seems a shame that they
allow 'All' for savesets but maybe not for directives?

Please post back if hard coding each file system does work.

Thanks

George

Pat DeCelie wrote:

George

Yes, the core files were backed up on a filesystem that is not the root FS.
However, this means I have to put the "+skip: core" for each filesystem.
That is a bummer. As for the length, I think it was 1024 chars until rev.
7 of NetWorker. So, that doesn't apply here. I'll check for jibberish.

Thanks,
Pat

--
Note: To sign off this list, send a "signoff networker" command via email
to listserv < at > listserv.temple.edu or visit the list's Web site at
http://listserv.temple.edu/archives/networker.html where you can
should be sent to stan < at > temple.edu

Note: To sign off this list, send a "signoff networker" command via email
to listserv < at > listserv.temple.edu or visit the list's Web site at
http://listserv.temple.edu/archives/networker.html where you can
should be sent to stan < at > temple.edu

Well, I'm not categorically stating that you have to. That was just my
surmise as to how it's little brain works. I mean, for example, if I
list say / under the savesets for a given client then is NetWorker gonna
back up just / or is it gonna back up all the file systems? Of course,
we all know it will descend no further than /; so, with the exception of
the mount point path names, any other file systems would not be captured
even though their mount points are under the root FS. I guess I was
thinking that directives would follow along the same logic?

But I would try it, and if doing it that way works ... natch. Should be
easy enough to test. As far as it being a bummer ... yes, it is. Might
be something in the man page, though, for uasm that might tell you how
to work around it?

The language in nsr(5) leads me to believe that it should traverse
filesystems. The language about the .nsr files is explicit that it does
do so. I see no reason that the server-fed directives would not also.


--
Darren Dunham ddunham < at > taos.com
Senior Technical Consultant TAOS http://www.taos.com/
Got some Dr Pepper? San Francisco, CA bay area
< This line left intentionally blank to confuse you. >

Note: To sign off this list, send a "signoff networker" command via email
to listserv < at > listserv.temple.edu or visit the list's Web site at
http://listserv.temple.edu/archives/networker.html where you can
should be sent to stan < at > temple.edu

Hmm ... if that's true then how would you skip certain files only under
/ ?

George

Darren Dunham wrote:

Well, I'm not categorically stating that you have to. That was just my
surmise as to how it's little brain works. I mean, for example, if I
list say / under the savesets for a given client then is NetWorker gonna
back up just / or is it gonna back up all the file systems? Of course,
we all know it will descend no further than /; so, with the exception of
the mount point path names, any other file systems would not be captured
even though their mount points are under the root FS. I guess I was
thinking that directives would follow along the same logic?

But I would try it, and if doing it that way works ... natch. Should be
easy enough to test. As far as it being a bummer ... yes, it is. Might
be something in the man page, though, for uasm that might tell you how
to work around it?

The language in nsr(5) leads me to believe that it should traverse
filesystems. The language about the .nsr files is explicit that it does
do so. I see no reason that the server-fed directives would not also.

--
Darren Dunham ddunham < at > taos.com
Senior Technical Consultant TAOS http://www.taos.com/
Got some Dr Pepper? San Francisco, CA bay area
< This line left intentionally blank to confuse you. >

--
Note: To sign off this list, send a "signoff networker" command via email
to listserv < at > listserv.temple.edu or visit the list's Web site at
http://listserv.temple.edu/archives/networker.html where you can
should be sent to stan < at > temple.edu

Note: To sign off this list, send a "signoff networker" command via email
to listserv < at > listserv.temple.edu or visit the list's Web site at
http://listserv.temple.edu/archives/networker.html where you can
should be sent to stan < at > temple.edu

On Tue, 5 Apr 2005, George Sinclair wrote:

Well, I'm not categorically stating that you have to. That was just my
surmise as to how it's little brain works. I mean, for example, if I
list say / under the savesets for a given client then is NetWorker gonna
back up just / or is it gonna back up all the file systems? Of course,
we all know it will descend no further than /; so, with the exception of
the mount point path names, any other file systems would not be captured
even though their mount points are under the root FS. I guess I was
thinking that directives would follow along the same logic?

I think the prudent thing to do is back up the core files. I was in a
situation a few years ago where a web server's disk crashed. Every bit of
data for that server's users was recovered, except for one user. Yup, that
user had the dumb luck to host a web site called "core". Since that
incident, I would just as soon back up those few core files than not. A
better solution is to find out what's generating core files on your server
and fix it if at all possible.

Note: To sign off this list, send a "signoff networker" command via email
to listserv < at > listserv.temple.edu or visit the list's Web site at
http://listserv.temple.edu/archives/networker.html where you can
should be sent to stan < at > temple.edu

In regard to: Re: [Networker] Core Files not skipped by Directives, George...:

Hmm ... if that's true then how would you skip certain files only under
/ ?

forget. See the man page that Darren referenced.

Tim
--
Tim Mooney mooney < at > dogbert.cc.ndsu.NoDak.edu
Information Technology Services (701) 231-1076 (Voice)
Room 242-J6, IACC Building (701) 231-8541 (Fax)
North Dakota State University, Fargo, ND 58105-5164

Note: To sign off this list, send a "signoff networker" command via email
to listserv < at > listserv.temple.edu or visit the list's Web site at
http://listserv.temple.edu/archives/networker.html where you can
should be sent to stan < at > temple.edu

In regard to: Re: [Networker] Core Files not skipped by Directives, George...:

Hmm ... if that's true then how would you skip certain files only under
/ ?

forget. See the man page that Darren referenced.

I've gone over that page, but I didn't see anything obvious.

I'm not certain how it would interpret something like...

<< / >>
skip: /file

Where the file has the / in front.

If I get a chance, I'll try to run that on the test machine later...

--
Darren Dunham ddunham < at > taos.com
Senior Technical Consultant TAOS http://www.taos.com/
Got some Dr Pepper? San Francisco, CA bay area
< This line left intentionally blank to confuse you. >

Note: To sign off this list, send a "signoff networker" command via email
to listserv < at > listserv.temple.edu or visit the list's Web site at
http://listserv.temple.edu/archives/networker.html where you can
should be sent to stan < at > temple.edu

Admins

Stan had said:
I think the prudent thing to do is back up the core files. I was in a
situation a few years ago where a web server's disk crashed. Every bit of
data for that server's users was recovered, except for one user. Yup, that
user had the dumb luck to host a web site called "core". Since that
incident, I would just as soon back up those few core files than not. A
better solution is to find out what's generating core files on your server
and fix it if at all possible.

"Ouch" about the core web site! At our site, we support several
environments, one of them being a Development environment. I agree it is
wise to determine what is causing the cores and fix it. Also, it may be
prudent to backup core files, but I see 1GB core files quite often and for
a 500 client backup domain, it probably adds up to a few terabytes of data
over the course of one year. I know there are some good jokes I could
insert here, but that is what it is. We have a 3 year retention for full
backups, so it can save a decent amout of tapes. One possibility is to
have a cron job find cores, email the owner of the core file and remove
ones older than 2 weeks.