Discussione:
IFS e SAVE
(troppo vecchio per rispondere)
Massimiliano Toninelli
2015-05-10 10:11:48 UTC
Permalink
Sto passando ad una macchina power 8 nuova;
ho circa 1,5 Tera di dati su IFS e dopo circa 950.000 oggetti salvati con un
semplice SAV xxxx mi dice che due oggetti non sono stati salvati;
ops...aveva lasciato aperto un doc word ed era alloccato.

Leggo un po' indietro e trovo una marea di msg CPD3775, ovvero che il file
non puo' essere salvato.


Informazioni aggiuntive sul messaggio

ID messaggio . . . . . : CPD3775 Severità . . . . . . . : 20
Tipo messaggio . . . . : Informativo
Data invio . . . . . . : 10/05/15 Ora invio . . . . . . : 09:31:39

Messaggio . . . : Non si può salvare o ripristinare
/portale/ton/clienti/claim/errore.pdf.
Causa . . . : Il sistema non consente il salvataggio o il ripristino
dell'oggetto /portale/ton/clienti/claim/errore.pdf.
Correzione. . . : Se si sta salvando o ripristinando l'intero sistema, non
è
richiesta nessuna correzione.
Altrimenti, eseguire una delle seguenti operazioni e ritentare la
richiesta:
-- Se /portale/ton/clienti/claim/errore.pdf è un indirizzario,
specificare
gli oggetti nell'indirizzario.
-- Se non è un indirizzario, specificare l'indirizzario contenente
l'oggetto.
Fine


Controllo meglio, e scopro che tantissimi file Pdf che non vengono salvati:

Non si può salvare o ripristinare /portale/ton/clienti/claim/chef.pdf
Non si può salvare o ripristinare /portale/ton/clienti/claim/chef.pdf.
Non si può salvare o ripristinare /portale/ton/clienti/claim/piatto.htm.
Non si può salvare o ripristinare /portale/ton/clienti/claim/piatto.pdf.
Non si può salvare o ripristinare /portale/ton/clienti/claim/15.pdf.

Vado a vederne uno a caso:

Visualizzazione attributi

Oggetto . . . . . : /portale/ton/clienti/claim/errore.pdf

Consentire scrittura durante
salvataggio . . . . . . . . . . . . : No
Salvabile . . . . . . . . . . . . . . : No


In pratica non riesco a salvare questi oggetti che sono su IFS in nessun
modo.

Per adesso ho fatto un cpy usando il QFileSvr.400, pero' è una cosa grave.
Chissà in questi anni durante i vari passaggi di macchina o save/restore ho
perso oggetti per colpa di quel piffero di parametro.

Qualcuno ne sa qualcosa?
Tia
--
--
Massimiliano
----------------------------------------
Chi sa fare, fa!
Chi non sa fare, insegna!
Chi non sa insegnare, insegna agli insegnanti!
e chi non sa insegnare agli insegnanti, fa politica.
http://www.toninelli.it
--
In CGIDEV2/AJAX/SSI
Forniture alberghiere http://www.nuovimagazzinirossi.it
CRPence
2015-05-10 17:36:18 UTC
Permalink
Post by Massimiliano Toninelli
Sto passando ad una macchina power 8 nuova;
ho circa 1,5 Tera di dati su IFS e dopo circa 950.000 oggetti
salvati con un semplice SAV xxxx mi dice che due oggetti non sono
stati salvati; ops...aveva lasciato aperto un doc word ed era
alloccato.
Leggo un po' indietro e trovo una marea di msg CPD3775, ovvero che
il file non puo' essere salvato.
Informazioni aggiuntive sul messaggio
ID messaggio . . : CPD3775 Severità . . . . : 20
Tipo messaggio . : Informativo
Data invio . . . : 10/05/15 Ora invio . . . : 09:31:39
Messaggio . . . : Non si può salvare o ripristinare
/portale/ton/clienti/claim/errore.pdf.
Causa . . . : Il sistema non consente il salvataggio o il
ripristino dell'oggetto /portale/ton/clienti/claim/errore.pdf.
Correzione. . . : Se si sta salvando o ripristinando l'intero
sistema, non è richiesta nessuna correzione. Altrimenti,
-- Se /portale/ton/clienti/claim/errore.pdf è un indirizzario,
specificare gli oggetti nell'indirizzario.
-- Se non è un indirizzario, specificare l'indirizzario
contenente l'oggetto.
Controllo meglio, e scopro che tantissimi file Pdf che non vengono
Non si può salvare o ripristinare /portale/.../claim/chef.pdf
Non si può salvare o ripristinare /portale/.../claim/chef.pdf.
Non si può salvare o ripristinare /portale/.../claim/piatto.htm.
Non si può salvare o ripristinare /portale/.../claim/piatto.pdf.
Non si può salvare o ripristinare /portale/.../claim/15.pdf.
Visualizzazione attributi
Oggetto . . . . . : /portale/ton/clienti/claim/errore.pdf
Consentire scrittura durante
salvataggio . . . . . . . . . . . . : No
Salvabile . . . . . . . . . . . . . . : No
Allow write during save . . . . . . . : No
Can be saved . . . . . . . . . . . . . : No

The "Can be saved" help text, unfortunately does not explain nor
allude to how or why the value might be set to "No"; one might suppose
to avoid being repetitive, but the general help for the Display
Attributes [for the option 8=Display Attributes from Work With Object
Links (WRKLNK)] also does not mention CHGATR anywhere. :-(
Post by Massimiliano Toninelli
In pratica non riesco a salvare questi oggetti che sono su IFS in
nessun modo.
Per adesso ho fatto un cpy usando il QFileSvr.400, pero' è una cosa
grave. Chissà in questi anni durante i vari passaggi di macchina o
save/restore ho perso oggetti per colpa di quel piffero di
parametro.
Qualcuno ne sa qualcosa?
The condition is logged as inconsequential [Sev=20 Informational]
because the file or the directory [of the file] had to have had the
Allow Save (*ALWSAV) attribute registered as *NO, _explicitly_, to
prevent the save. Setting that attribute is typical for log files [both
temporal and often locked] and data that would be backed-up elsewhere
[i.e. data that is on local disk for convenience, perhaps cached, but
the primary repository is elsewhere, or the data is purely temporal and
insignificant for effecting a proper full system recovery. Thus why the
message talks about a full system save, as the Backup and Recovery (B&R)
plan should include separate SAV activity for files in the directories
that were explicitly excluded by the effective /disallow save/
attribute, *if* the files are required for effecting a full system
recovery as part of B&R.

See the Change Attribute (CHGATR) command for the Attribute (ATR)
parameter for which the New Value (VALUE) parameter can specify either
*YES or *NO for the *ALWSAV attribute. See the help text [or the
corresponding docs in the KnowledgeCenter] for the ATR parameter.
--
Regards, Chuck
Massimiliano Toninelli
2015-05-12 07:43:52 UTC
Permalink
Post by CRPence
Post by Massimiliano Toninelli
Sto passando ad una macchina power 8 nuova;
ho circa 1,5 Tera di dati su IFS e dopo circa 950.000 oggetti
salvati con un semplice SAV xxxx mi dice che due oggetti non sono
stati salvati; ops...aveva lasciato aperto un doc word ed era
alloccato.
Leggo un po' indietro e trovo una marea di msg CPD3775, ovvero che
il file non puo' essere salvato.
Informazioni aggiuntive sul messaggio
ID messaggio . . : CPD3775 Severità . . . . : 20
Tipo messaggio . : Informativo
Data invio . . . : 10/05/15 Ora invio . . . : 09:31:39
Messaggio . . . : Non si può salvare o ripristinare
/portale/ton/clienti/claim/errore.pdf.
Causa . . . : Il sistema non consente il salvataggio o il
ripristino dell'oggetto /portale/ton/clienti/claim/errore.pdf.
Correzione. . . : Se si sta salvando o ripristinando l'intero
sistema, non è richiesta nessuna correzione. Altrimenti,
-- Se /portale/ton/clienti/claim/errore.pdf è un indirizzario,
specificare gli oggetti nell'indirizzario.
-- Se non è un indirizzario, specificare l'indirizzario
contenente l'oggetto.
Controllo meglio, e scopro che tantissimi file Pdf che non vengono
Non si può salvare o ripristinare /portale/.../claim/chef.pdf
Non si può salvare o ripristinare /portale/.../claim/chef.pdf.
Non si può salvare o ripristinare /portale/.../claim/piatto.htm.
Non si può salvare o ripristinare /portale/.../claim/piatto.pdf.
Non si può salvare o ripristinare /portale/.../claim/15.pdf.
Visualizzazione attributi
Oggetto . . . . . : /portale/ton/clienti/claim/errore.pdf
Consentire scrittura durante
salvataggio . . . . . . . . . . . . : No
Salvabile . . . . . . . . . . . . . . : No
Allow write during save . . . . . . . : No
Can be saved . . . . . . . . . . . . . : No
The "Can be saved" help text, unfortunately does not explain nor allude
to how or why the value might be set to "No"; one might suppose to avoid
being repetitive, but the general help for the Display Attributes [for the
option 8=Display Attributes from Work With Object Links (WRKLNK)] also
does not mention CHGATR anywhere. :-(
Post by Massimiliano Toninelli
In pratica non riesco a salvare questi oggetti che sono su IFS in
nessun modo.
Per adesso ho fatto un cpy usando il QFileSvr.400, pero' è una cosa
grave. Chissà in questi anni durante i vari passaggi di macchina o
save/restore ho perso oggetti per colpa di quel piffero di
parametro.
Qualcuno ne sa qualcosa?
The condition is logged as inconsequential [Sev=20 Informational]
because the file or the directory [of the file] had to have had the Allow
Save (*ALWSAV) attribute registered as *NO, _explicitly_, to prevent the
save. Setting that attribute is typical for log files [both temporal and
often locked] and data that would be backed-up elsewhere [i.e. data that
is on local disk for convenience, perhaps cached, but the primary
repository is elsewhere, or the data is purely temporal and insignificant
for effecting a proper full system recovery. Thus why the message talks
about a full system save, as the Backup and Recovery (B&R) plan should
include separate SAV activity for files in the directories that were
explicitly excluded by the effective /disallow save/ attribute, *if* the
files are required for effecting a full system recovery as part of B&R.
See the Change Attribute (CHGATR) command for the Attribute (ATR)
parameter for which the New Value (VALUE) parameter can specify either
*YES or *NO for the *ALWSAV attribute. See the help text [or the
corresponding docs in the KnowledgeCenter] for the ATR parameter.
Hi Chuck,
I think is not correct that I don't receive a msg alert ending the command
save.
I had a lot of migration of servers and I don't kwow how many obj I lost.
I don't know why this attribute has the value NO; I use 5721 Infoprint
server for create invoices.
At the moment, I realize to have lost a lot of invoices.
But also the command SAV doesn't have a parm which include o not the object
not save.
If I want to save *ALLSYS, I want all, not somenthing..
Sorry... but I am very incazzed !!! :-D
Ciao
Max
CRPence
2015-05-12 13:30:07 UTC
Permalink
Post by Massimiliano Toninelli
I think is not correct that I don't receive a msg alert ending the
command save.
The [dis]Allow Save feature is intentionally /silent/; intended not
to _fail_ the save, because the files modified to have that designation
[or placed in directories with that designation] are presumed
necessarily to have been set explicitly to be excluded by SAV [or were
explicitly chosen to be placed in a directory that should be excluded by
SAV].

Been too long since I did any SAV work; is there possibly in the
completion message, a summary of object counts; one that might include
those excluded for the specific reason being that attribute setting?

The situation would be much like saving something that is *STGFREE
such that performing the save is uninhibited and perfectly acceptable
processing, yet the ramifications could be highly negative if the proper
planning for B&R did not ensure the previously /freed/ storage exists
and is available on preserved media. The same story for the objects
excluded by *ALWSAV=no.
Post by Massimiliano Toninelli
I had a lot of migration of servers and I don't know how many obj
I lost.
Ouch. I can only commiserate with you in that regard.

Hopefully not sounding too crass... but there are the CPD3775
messages logged to the joblog of the save, and AFaIK similar information
would be logged in the OUTPUT() specifications. Additionally the
DSPJOBLOG does have an OutFile specification for which a simple query
can find messages that may be of interest, or excluding those which are
accepted as recording only a positive effect. There would also be no
indication that the object is saved, recorded with those objects; no
last-saved information. Anyhow, no matter what information is generated
[or info conspicuous by absence, as in last-saved details], unless there
is somebody to look, that information goes unnoticed, and that can be
problematic. :-(
Post by Massimiliano Toninelli
I don't know why this attribute has the value NO; I use 5721
Infoprint server for create invoices.
At the moment, I realize to have lost a lot of invoices.
I suppose just like I had suggested in my prior reply, the feature
creating those objects [and\or directories] probably deems these to be
/copies/ of data that would be available elsewhere "and insignificant
for effecting a proper full system recovery." I would expect the
documentation for that software would have the details explaining that
attribute setting as the effect, and the means to override that [as the
apparent default] behavior. And if not, probably an upcoming project
will be to /correct/ any such designations after any new directories
and\or files are created by that feature; possibly in conjunction with
asking the software providers to enhance the feature to provide that
override capability so as to avoid having to reset the objects.
Post by Massimiliano Toninelli
But also the command SAV doesn't have a parm which include o not the
object not save.
IIRC the SAV will still function to save files in a directory marked
with the disallow-save; i.e. the save of the directory with the files
that exhibits the exclusion. But that is why, for a B&R plan, the
full-save must be *supplemented* by additional SAV requests to backup
what was purposely excluded by the *ALWSAV settings. That should be
documented in the B&R planning; used to be, there was an Advanced Backup
and Recovery Guide, but I am not sure if that is incorporated into the
KnowledgeCenter or ¿what?.
Post by Massimiliano Toninelli
If I want to save *ALLSYS, I want all, not something..
I suppose submitting a requirement to enable the *ALWSAV attribute to
support a new value such as *ONLYFULLSAVE; i.e. a value other than *YES
and *NO, for which the save should be attempted rather than implicitly
omitted. Asking for the SAV to offer an override to the disallow-save
feature seems very unlikely, because the *NO *ALWSAV setting is quite an
extensively utilized capability for which an actual save is truly
undesirable. Asking that the SAV offer a parameter that would suggest
the CPD3775 be treated as a diagnostic might be more workable because,
changing the behavior except perhaps to be a little more /informative/
than just the Informational message CPD3775, is likely all that would be
offered; most likely one would be directed to see the OUTPUT(),
presuming that effect is logged there. There are just too many good
reasons the *NO setting should remain silent, despite any thought that
All=All.

FWiW, AFaIK, the BRMS feature produces a report showing what will not
be saved, and that would include directories designated as disallowed
for save.?
--
Regards, Chuck
CRPence
2015-05-12 14:35:57 UTC
Permalink
FWiW: A message thread [topic still active] that discusses recovery
planning and testing and a mention of *ALWSAV attribute; a quote worth
mention, is the suggestion to "Plan the recovery first, and then do the
saves to support it" rather than doing\planning a save only to learn
that recovery is not possible from what was saved.
<http://archive.midrange.com/midrange-l/201505/threads.html#00226>

BTW, and OT; mostly just a public _rant_, but informational if others
are similarly affected: My browser has an issue with the formatting of
some of the messages; the scroll-bar is in the message instead of the
browser page, making viewing of the content almost unusable. I recall
the archive owner said there were changes, but I had not noticed the
SNAFU with the latest announced change. I just reviewed the same
content via the gmane.org archive to see if that presentation was [more]
usable, and found the content is much cleaner; although I recall that
interface has some different issues, I just do not recall what, but
something to do with replacing valid strings in content with what seems
like nonsense, at least until the pattern is recognized and what the
replacement intends. In either case, if problems viewing, the [albeit
crappy] option exists to ask the browser to /view-page-source/ and then
copy the textual content elsewhere, e.g. an editor, possibly enabling
easier viewing, thus avoiding the /formatting-gone-amuck/ with HTML
tooling rather than the archives just showing the plain-text copy of
archived messages.

The thread started 11 May 20:46 2015 by LRoberts with "Subject: BRMS
Recovery Script Questions":
<http://news.gmane.org/gmane.comp.hardware.ibm.midrange>
--
Regards, Chuck
Massimiliano Toninelli
2015-05-13 13:31:56 UTC
Permalink
Post by CRPence
FWiW: A message thread [topic still active] that discusses recovery
planning and testing and a mention of *ALWSAV attribute; a quote worth
mention, is the suggestion to "Plan the recovery first, and then do the
saves to support it" rather than doing\planning a save only to learn that
recovery is not possible from what was saved.
<http://archive.midrange.com/midrange-l/201505/threads.html#00226>
BTW, and OT; mostly just a public _rant_, but informational if others
are similarly affected: My browser has an issue with the formatting of
some of the messages; the scroll-bar is in the message instead of the
browser page, making viewing of the content almost unusable. I recall the
archive owner said there were changes, but I had not noticed the SNAFU
with the latest announced change. I just reviewed the same content via
the gmane.org archive to see if that presentation was [more] usable, and
found the content is much cleaner; although I recall that interface has
some different issues, I just do not recall what, but something to do with
replacing valid strings in content with what seems like nonsense, at least
until the pattern is recognized and what the replacement intends. In
either case, if problems viewing, the [albeit crappy] option exists to ask
the browser to /view-page-source/ and then copy the textual content
elsewhere, e.g. an editor, possibly enabling easier viewing, thus avoiding
the /formatting-gone-amuck/ with HTML tooling rather than the archives
just showing the plain-text copy of archived messages.
The thread started 11 May 20:46 2015 by LRoberts with "Subject: BRMS
<http://news.gmane.org/gmane.comp.hardware.ibm.midrange>
Hi Chuck thank you for the explanation end my visual point of what you
written is changed...
but...I don'k know why the stmfs has changed with that parm.
Tnx
CRPence
2015-05-13 18:20:10 UTC
Permalink
<<SNIP>>
but...I don'k know why the stmfs has changed with that parm.
If nobody [no user] had changed those Stream Files (STMF), then the
application that creates them most likely did. To see when they get
changed, turn on the appropriate auditing to track; likely after
creation of the file, there will be an effective ChgAtr request being
made by the application [apparently the "5721 Infoprint server"] that
created the invoices as STMF. If such auditing is already active, then
what exists of the online logs can be searched for that change activity.

As I suggested in a prior reply, the software feature creating those
files should probably have an option to configure if the files should
*not* be created that way; i.e. give a method to tell the application to
create the files as Allow-Save (*ALWSAV=*YES) instead of create the file
and then change them to have the DisAllow-Save (*ALWSAV=*NO) attribute
setting.

The OS SAV feature is just _honoring_ the setting; and apparently too
silently for your tastes :-( Do please check however, if the OUTPUT()
capabilities of the SAV do divulge the CPD3775 effect. If they are not
included, perhaps that would be a good enhancement request, to add a
category of for object excluded per Disallow Save specification?; on
v5r3, there seems to be nothing logged about those, neither for
Information Type (INFTYPE) of *ALL or *ERR.
--
Regards, Chuck
Dr.UgoGagliardelli
2015-05-12 14:19:16 UTC
Permalink
Il 12.05.2015 09.43, Massimiliano Toninelli ha scritto:
[...]
Post by Massimiliano Toninelli
Sorry... but I am very incazzed !!! :-D
Sistemista fai da te? Ahi, ahi, ahi...
:-)
Massimiliano Toninelli
2015-05-13 11:54:13 UTC
Permalink
Post by Dr.UgoGagliardelli
[...]
Post by Massimiliano Toninelli
Sorry... but I am very incazzed !!! :-D
Sistemista fai da te? Ahi, ahi, ahi...
:-)
A dottò: :-ppppppp

:-)
Obelix
2015-05-13 09:26:40 UTC
Permalink
Post by Massimiliano Toninelli
I think is not correct that I don't receive a msg alert ending the command
save.
Uh ??? Sorry, buddy: if you (or anyone else..) said to the system "i
don't want you to save this object' why should the system complaint
diring the save??

I would be incazzed if it does. As Chuck said, this should be set for
temporary or low-insteresting object who you doesn't need/want to save.

The fault IMHO is in the proc that set that bit to *NO, not on the save
that works correctly.
Post by Massimiliano Toninelli
Post by Massimiliano Toninelli
But also the command SAV doesn't have a parm which include o not the
object not save.
Again: if i said to system to not save some object, WTH sould be saying
to system to save ???
Post by Massimiliano Toninelli
Post by Massimiliano Toninelli
If I want to save *ALLSYS, I want all, not somenthing..
*ALLSYS still doesn't save everything, you know?? apart from the
*alwsav() parm
Massimiliano Toninelli
2015-05-13 13:33:49 UTC
Permalink
Post by Massimiliano Toninelli
I think is not correct that I don't receive a msg alert ending the command
save.
Uh ??? Sorry, buddy: if you (or anyone else..) said to the system "i don't
want you to save this object' why should the system complaint diring the
save??
I would be incazzed if it does. As Chuck said, this should be set for
temporary or low-insteresting object who you doesn't need/want to save.
The fault IMHO is in the proc that set that bit to *NO, not on the save
that works correctly.
Post by Massimiliano Toninelli
Post by Massimiliano Toninelli
But also the command SAV doesn't have a parm which include o not the
object not save.
Again: if i said to system to not save some object, WTH sould be saying to
system to save ???
Post by Massimiliano Toninelli
Post by Massimiliano Toninelli
If I want to save *ALLSYS, I want all, not somenthing..
*ALLSYS still doesn't save everything, you know?? apart from the *alwsav()
parm
I'am the boss!!!! If I tell to my system to save all, it must save all!!!!!
:-)
Otherwise, put a parm in the savsys which asks me if I want to save the
object non salvable!!!!
Obelix
2015-05-13 15:08:39 UTC
Permalink
Post by Massimiliano Toninelli
I'am the boss!!!! If I tell to my system to save all, it must save all!!!!!
:-)
There's a Bigger boss in IBm :)
Post by Massimiliano Toninelli
Otherwise, put a parm in the savsys which asks me if I want to save the
object non salvable!!!!
What about ask why your app put that flag to *NO??? default value is *YES...
Massimiliano Toninelli
2015-05-13 15:22:47 UTC
Permalink
Post by Obelix
Post by Massimiliano Toninelli
I'am the boss!!!! If I tell to my system to save all, it must save all!!!!!
:-)
There's a Bigger boss in IBm :)
Azz!!! Ubi major, minor cessat!
Post by Obelix
Post by Massimiliano Toninelli
Otherwise, put a parm in the savsys which asks me if I want to save the
object non salvable!!!!
What about ask why your app put that flag to *NO??? default value is *YES...
Ecche ne saccie io!!! Ajje trovato cussì!
Loading...