RE: Back ups with High Availability Web Server

We had a similar issue several weeks ago.
Over the weekend, we had multiple issues causing the library to run out of scratch tapes. (Friday's scratch tapes didn't get loaded because IO door was open, another scratch tape went into error status)
Sunday's nights save went to a 2nd tape while saving the IFS, scratch tape wasn't available, save went to MSGW.
Some, not all, Monday morning production jobs that needed IFS access had failures.
I had to cancel the save immediately, didn't have time to see what was locked.
The one failure was a CPYTOIMPF, path below, error was a 3029.

I'm building a new monitor into the save, if not done by 4:00 AM, save will end.

*NONE Command 10/28/13 07:45:02.125617 QCADRV QSYS 03B9 NDACLP03 CMPGMS 002A
Message . . . . : 1101 - CPYTOIMPF FROMFILE(CMFILES/NDAPHY01)
TOSTMF('/home/SFTPXFER/NDADirectory/dirFiles/PALMDIR') MBROPT(*REPLACE)
FROMCCSID(37) STMFCODPAG(*PCASCII) STMFAUT(*INDIR) RCDDLM(*CRLF)
DTAFMT(*FIXED) STRDLM(*NONE) NULLIND(*NO) ORDERBY(*ARRIVAL)
CPFA0D4 Diagnostic 40 10/28/13 07:45:02.210174 QCPEXPRT QSYS *STMT NDACLP03 CMPGMS 002A
From module . . . . . . . . : QCPEXPRT
From procedure . . . . . . : Send_type_msg
Statement . . . . . . . . . : 17
Message . . . . : File system error occurred. Error number 3029.

-----Original Message-----
From: midrange-l-bounces@xxxxxxxxxxxx [mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of Bradley Stone
Sent: Wednesday, November 13, 2013 11:37 AM
To: Midrange Systems Technical Discussion
Subject: Re: Back ups with High Availability Web Server

Paul,

I've only seen it in cases where a save is either saving files in a directory, or the tape is in MSGW status while saving a file in a directory.

Other applications seemed to be locked out from any directory or file in the path. I've seen it many times with our software that works with the IFS a lot for stream files.

When I get the call with the locking error the first questions I ask are, "when does it happen" and "when do you do backups" and "is there a MSGW on a save right now"?

Normally the answers are "early in the morning", "they run all night into
the morning", and "oops, yes". :)

Brad


On Wed, Nov 13, 2013 at 10:32 AM, Steinmetz, Paul <PSteinmetz@xxxxxxxxxx>wrote:

Brad,

I'd like to confirm something you said.

First, when saving an IFS object, the entire directory is locked until
the save is complete.
So if an application was trying to access that same directory, (but
not necessarily the same object), that app would probably fail because
of a lock.

Secondly, if an app has an IFS object "in use" then the save tries to
save that directory and object, only that "locked" object will be
skipped, but the rest of the directory would be saved.

Is this correct?

I have issues with both, I am trying to improve the save, but also
trying to guarantee that the save will not cause an app to fail.

Paul



-----Original Message-----
From: midrange-l-bounces@xxxxxxxxxxxx [mailto:
midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of Bradley Stone
Sent: Tuesday, November 12, 2013 11:10 PM
To: Midrange Systems Technical Discussion
Subject: Re: Back ups with High Availability Web Server

We are talking about the majority of the data, and what's taking so
long, are the IFS objects. Not really any data in QSYS.LIB.

I only mention this because there seems to be a little confusion, and
I've experienced in the past when they're saving an IFS object, that
entire directory is locked until it's done. And locked down hard. :)

Brad


On Tue, Nov 12, 2013 at 7:04 PM, Charles Wilt <charles.wilt@xxxxxxxxx

wrote:

I don't see much difference between VSS and IBM i's Save While Active...

Except that SWA has been available since at least 1997 or
1996...probably longer.

The problem is too many shops don't bother to use it.

Charles


On Tue, Nov 12, 2013 at 4:25 PM, Matt Olson <Matt.Olson@xxxxxxxx> wrote:

First question is have you ever restored one of these? How does
the operating system come up when you do that?

Yes, comes up in recovery mode the first bootup, subsequent
bootups are normal.

Second is what are you doing with those servers? Are these
hosting

large

Oracle or MSSQL Databases or are they file and print servers? Or
are they just AD members that are used for remote desktop for example.

100GB database servers, print servers, file servers, web servers,
you

name

it. Everything in windows these days have volume shadow copy

providers.

IBM i really needs something like Windows volume shadow copy
service so you can achieve 100% up time during backups.

This is the only reason why I see AS400 shops are down
periodically, had the database been hosted on one of a windows box
we would have much

better

uptime.

-----Original Message-----
From: DrFranken [mailto:midrange@xxxxxxxxxxxx]
Sent: Tuesday, November 12, 2013 3:21 PM
To: Midrange Systems Technical Discussion
Subject: Re: Back ups with High Availability Web Server

Matt,

First question is have you ever restored one of these? How does
the operating system come up when you do that?

Second is what are you doing with those servers? Are these
hosting

large

Oracle or MSSQL Databases or are they file and print servers? Or
are they just AD members that are used for remote desktop for example.

There is a large difference between a transaction oriented
system and a system hosting many flat (or 'stream') files.

- Larry "DrFranken" Bolhuis

www.frankeni.com
www.iDevCloud.com
www.iInTheCloud.com

On 11/12/2013 4:06 PM, Matt Olson wrote:

IBM really needs to fix this issue. Put in a Design Change
Request to

allow for online backups like you can do with Windows. We do full
(bare
metal) backups of our windows machines all the time, no downtime

required.

--
This is the Midrange Systems Technical Discussion (MIDRANGE-L)
mailing list To post a message email: MIDRANGE-L@xxxxxxxxxxxx To
subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx Before posting, please
take a moment to review the archives at

http://archive.midrange.com/midrange-l.

--
This is the Midrange Systems Technical Discussion (MIDRANGE-L)
mailing

list

To post a message email: MIDRANGE-L@xxxxxxxxxxxx To subscribe,
unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx Before posting, please
take a moment to review the archives at
http://archive.midrange.com/midrange-l.

--
This is the Midrange Systems Technical Discussion (MIDRANGE-L)
mailing list To post a message email: MIDRANGE-L@xxxxxxxxxxxx To
subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx Before posting, please
take a moment to review the archives at
http://archive.midrange.com/midrange-l.

--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing
list To post a message email: MIDRANGE-L@xxxxxxxxxxxx To subscribe,
unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx Before posting, please take
a moment to review the archives at http://archive.midrange.com/midrange-l.

--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing
list To post a message email: MIDRANGE-L@xxxxxxxxxxxx To subscribe,
unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx Before posting, please take
a moment to review the archives at
http://archive.midrange.com/midrange-l.

--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list To post a message email: MIDRANGE-L@xxxxxxxxxxxx To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx Before posting, please take a moment to review the archives at http://archive.midrange.com/midrange-l.