Onbar and Storage Manager Backup to different server mpossible?

Hi guys. We have 2 IBM P570 servers, one production one as DR. Both machine run IDS 9.4.FC6 and are exactly the same EXCEPT for the hostname. Policy of the datacenter here we have the production server makes and the fact that the program running on it doesn't usetransaction logging makes it impossible to use HDR. Neither can we use mirror because the DR isn't physically in the same building and the return time is 30minutes. This is a 24/7 system. So we use the Storage manager and Onbar. All is setup by the book (Onbar and Storage manager manual). The problem is that the DR onbar doesn't allow the backup to be loaded into the instance because the administrator on the DRmachine isn't known on the production machine (not in the remote access list). We have tried, and tried, and tried. We can not get beyond this error. it seems to be directly connected to the difference in the hostname, but according to the manuals it should. Anyone any idea? Art Kagel, Jonathan Lefflermaybe? i'm stomped. IBM tells me 'hmmm, it may not be possible' which frankly isses me off mightily. Robert A. Reissaus Risdi (Amsterdam, Brussel, Zurich, Chicago) r.a.reissaus@risdi.com

Robert A. Reissaus wrote: > We have 2 IBM P570 servers, one production one as DR. Both machine run > IDS 9.4.FC6 and are exactly the same EXCEPT for the hostname. DR? What does that mean? (I can guess, but I don't know.) > Policy of the datacenter here we have the production server makes and > the fact that the program running on it doesn't usetransaction logging > makes it impossible to use HDR. Neither can we use mirror because the > DR isn't physically in the same building and the return time is > 30minutes. This is a 24/7 system. 'Return time'? What does that mean? (I'm not even sure I can guess this one.) And not using a transaction log with IDS is pretty silly - you should normally use a transaction log unless there are really strong mitigating circumstances (usually, a third party supplier who insists on not using a transaction log - but I'd look askance at an application that doesn't use transactions anyway). (SE is a different story.) > So we use the Storage manager and Onbar. All is setup by the book > (Onbar and Storage manager manual). The problem is that the DR onbar > doesn't allow the backup to be loaded into the instance because the > administrator on the DRmachine isn't known on the production machine > (not in the remote access list). Which storage manager? ISM - the Informix Storage Manager? That is a cut down version of a commercial product (Legato Networker, IIRC), and has some limitations - quite deliberate limitations. You are encouraged to use a separately purchased storage manager. That said, within the limitations, ISM is fully functional. I'm not clear what your problem is - which administrator on the DR machine? Which remote access list? What are the exact error messages? Which files have you copied across? > We have tried, and tried, and tried. We can not get beyond this error. > it seems to be directly connected to the difference in the hostname, > but according to the manuals it should. > > Anyone any idea? > Art Kagel, Jonathan Lefflermaybe? i'm stomped. IBM tells me 'hmmm, it > may not be possible' which frankly isses me off mightily. I'm not the BAR expert. So, you may be talking using perfectly sensible terminology, but there are a lot of terms I don't recognize in your description. You've not provided the exact error messages you are seeing. You've not explained exactly what steps you're taking. You've not explained whether you're trying to recover the target system from archives every day - or any other interval. So, as yet, I don't understand the problem, and cannot offer a solution. -- Jonathan Leffler #include <disclaimer.h> Email: jleffler@earthlink.net, jleffler@us.ibm.com Guardian of DBD::Informix v2005.01 -- http://dbi.perl.org/

-- administrator on the DRmachine isn't known on the production machine ahum ism_add -admin <someuser> ??? Sounds silly; you may have a look at setting up ism for HDR your answer should /could be there. What you need i guess is: the catalog from ism and the ixbar.<servernum> file in $INFOMIXDIR/etc/ together with onconfig.... getting the catalog means you must copy this to tape (check out onbar script and manual) if its on your tape then you can get it of using ism_catalog check the manual for it; it is some time ago i used it; one of them is reading your whole tape and may give you grey hair when it's finished (takes a long while...) Superboer.

Robert A. Reissaus wrote: > Hi guys. > We have 2 IBM P570 servers, one production one as DR. Both machine run > IDS 9.4.FC6 and are exactly the same EXCEPT for the hostname. > > Policy of the datacenter here we have the production server makes and > the fact that the program running on it doesn't usetransaction logging > makes it impossible to use HDR. Doesn't matter. If you change the database to (buffered) logging mode, all inserts, updates and deletes are logged and you can use HDR. When the application don't use BEGIN WORK and COMMIT/ROLLBACK WORK every single insert/update/delete is considered committed and written to the transaction log.

[SNIP] Jonathan Leffler <jleffler@earthlink.net> wrote in message news:<42B0F689.6070107@earthlink.net>... [SNIP] DR stands for Disaster Recovery. Since I came across it again and again and again in the past 15 years a well enough known abbreviation I would say. Return time is the time elapsed between sending data from the source to the target and getting the result code ( ack or nack) back from the target to the source. Something I believe is also a well known topic. Yes of course ISM. I know it is 'small version of Legato' but that isn't important here now is it? It comes bundledwith the engine and Onbar and should do what it says in the manuals. And it doesn't when doing a remote restore. What file have copied across? I said it is EXACTLY the same as the source except for the Hostname didn't I? If you are not a onbar expert, even when I asked for a response, a simple 'I am not an onbar expert so I can't help you here' would have been more than enough. The rest is totally unwarrented. Anyone else have something usefull to say? Nonetheless, Thanksfor your trouble. Robert A. Reissaus Ibm Informix consultants. Risdi (r) (tm) (Amsterdam, Brussels, Zurich, Chicago)

You asked for help and you had someone try to help you in good faith.

Claus Samuelsen wrote: > Robert A. Reissaus wrote: >> We have 2 IBM P570 servers, one production one as DR. Both machine run >> IDS 9.4.FC6 and are exactly the same EXCEPT for the hostname. >> >> Policy of the datacenter here we have the production server makes and >> the fact that the program running on it doesn't usetransaction logging >> makes it impossible to use HDR. > > Doesn't matter. If you change the database to (buffered) logging mode, > all inserts, updates and deletes are logged and you can use HDR. When > the application don't use BEGIN WORK and COMMIT/ROLLBACK WORK every > single insert/update/delete is considered committed and written to the > transaction log. Nearly - there's just the little detail that a cursor FOR UPDATE must be opened inside a transaction. Also, in an unlogged database, you must explicitly unlock any tables you lock; in a logged database, you must lock tables inside a transaction and you cannot explicitly unlock any tables you lock. -- Jonathan Leffler #include <disclaimer.h> Email: jleffler@earthlink.net, jleffler@us.ibm.com Guardian of DBD::Informix v2005.01 -- http://dbi.perl.org/

Robert A. Reissaus wrote: > [SNIP] > Jonathan Leffler <jleffler@earthlink.net> wrote: > [SNIP] > > DR stands for Disaster Recovery. Since I came across it again and > again and again in the past 15 years a well enough known abbreviation > I would say. OK - I don't remember coming across the term used for that purpose (certainly not sufficiently routinely for it to have sunk in), so I've been asleep. Must be the Rip van Winkle effect - another 21 years and I'll wake up. > Return time is the time elapsed between sending data from the source > to the target and getting the result code ( ack or nack) back from the > target to the source. Something I believe is also a well known topic. Again, I've not seen that as the description - again, I must have been dozing at all the wrong points. > Yes of course ISM. I know it is 'small version of Legato' but that > isn't important here now is it? It comes bundled with the engine and > Onbar and should do what it says in the manuals. And it doesn't when > doing a remote restore. OK; so you've contacted IBM Tech Support and asked for a bug to be created, clear stating exactly where there is a discrepancy between the documentation and the actual product behaviour? > What file have copied across? I said it is EXACTLY the same as the > source except for the Hostname didn't I? The sqlhosts file is the same? The dbservernames are the same? If the sqlhosts files are the same, are you using localhost as the machine name? Or is that what you meant by 'except for the hostname'? Did you copy the various BAR-related log files - which go by names such as ixbar*, I believe? Did you make all the necessary changes (if there are any) in the ON-Bar configuration files. > If you are not a onbar expert, even when I asked for a response, a > simple 'I am not an onbar expert so I can't help you here' would have > been more than enough. The rest is totally unwarrented. I am sorry that my attempt to understand your problem annoyed you. I was merely trying to obtain enough information from you to have a chance of helping you. I knew that my answer would, of itself, be of little value - but you did call upon me when, unasked, I'd have kept quiet, knowing that I have little of use to say. > Anyone else have something usefull to say? I suspect Martin gave you (at least some of) the extra information that you need. -- Jonathan Leffler #include <disclaimer.h> Email: jleffler@earthlink.net, jleffler@us.ibm.com Guardian of DBD::Informix v2005.01 -- http://dbi.perl.org/

> You asked for help and you had someone try to help you in good faith. > From your response I would never give you any assistance in the future. me neither. specially since this is Dutch and comes from Amsterdam. and i have a good memory for these things. Superboer. Doug McAllister schreef: > You asked for help and you had someone try to help you in good faith. > From your response I would never give you any assistance in the future. > > "Robert A. Reissaus" <r.a.reissaus@risdi.com> wrote in message > news:87ad5966.0506160224.c78ca29@posting.google.com... > > [SNIP] > > Jonathan Leffler <jleffler@earthlink.net> wrote in message > news:<42B0F689.6070107@earthlink.net>... > > [SNIP] > > > > DR stands for Disaster Recovery. Since I came across it again and > > again and again in the past 15 years a well enough known abbreviation > > I would say. > > > > Return time is the time elapsed between sending data from the source > > to the target and getting the result code ( ack or nack) back from the > > target to the source. Something I believe is also a well known topic. > > > > Yes of course ISM. I know it is 'small version of Legato' but that > > isn't important here now is it? It comes bundledwith the engine and > > Onbar and should do what it says in the manuals. And it doesn't when > > doing a remote restore. > > > > What file have copied across? I said it is EXACTLY the same as the > > source except for the Hostname didn't I? > > > > If you are not a onbar expert, even when I asked for a response, a > > simple 'I am not an onbar expert so I can't help you here' would have > > been more than enough. The rest is totally unwarrented. > > > > Anyone else have something usefull to say? > > > > Nonetheless, > > Thanksfor your trouble. > > Robert A. Reissaus > > Ibm Informix consultants. > > Risdi (r) (tm) (Amsterdam, Brussels, Zurich, Chicago)

Jonathan Leffler <jleffler@earthlink.net> wrote in message news:<q_sse.5485$jX6.3156@newsread2.news.pas.earthlink.net>... > > OK; so you've contacted IBM Tech Support and asked for a bug to be > created, clear stating exactly where there is a discrepancy between the > documentation and the actual product behaviour? Yep I did. In a 2 page point by point explanation. To them I said what I said here: 'the machines are EXACTLY the same except for the 'hostname' > > > What file have copied across? I said it is EXACTLY the same as the > > source except for the Hostname didn't I? > > The sqlhosts file is the same? The dbservernames are the same? If the > sqlhosts files are the same, are you using localhost as the machine > name? Or is that what you meant by 'except for the hostname'? Yes, That is EXACTLY what I meant. And of course the 3rd field in sqlhosts has changed. > Did you > copy the various BAR-related log files - which go by names such as > ixbar*, I believe? Did you make all the necessary changes (if there are > any) in the ON-Bar configuration files. Yes we did. The Disaster Recovery (as Obnoxio pointed out it could also be Data Replication) machine is a clone of the production, with the same hardware components, environment, etc,etc, etc. And yes , we did copy / change all the files (uptill and including the ism_startup). As I said 'BY TE BOOK' You see Jonathan I do RTFM. Every time again since my clients have versions going from 2.0 SE (!) till 10 and because I am not arrogant enough to think I know it all (even after 18years). > > > If you are not a onbar expert, even when I asked for a response, a > > simple 'I am not an onbar expert so I can't help you here' would have > > been more than enough. The rest is totally unwarrented. > > I am sorry that my attempt to understand your problem annoyed you. I > was merely trying to obtain enough information from you to have a > chance of helping you. I knew that my answer would, of itself, be of > little value - but you did call upon me when, unasked, I'd have kept > quiet, knowing that I have little of use to say. > Well, I did go off here a bit didn't I. my apologies. But after hearing 'did you do this, that, here and there for almost a week from all kinds of 'specialists' that also missed the meaning of the word 'exact' and the phrase 'by the book' I got just a little, a well, 6 letters with starting with 'p'. So I do apologize. > > I suspect Martin gave you (at least some of) the extra information that > you need. Well,no. He just told me what i got 'from the book' (:-)) For now we will use good old ontape.At least I could load the data onto the other instance without a problem. I think I'll start digging in the ism file to see what I can find. According to IBM it should work, and I remember it did in version 7.3 or 8 so why it doesn't here I don't now. Can't wait for an answer. Only thing I need is to find a neat way to have ontape not needing attention. Next week i'll build myself a nice small instance all for me on the 2 different machines and start digging. If I find out why i'll let you know. Robert A. Reissaus IBM/Informix Consultants (Amsterdam, Brussel, Chicago, Zurich)

{SNIP] "Superboer" <superboer7@planet.nl> wrote in message news:<1118990819.498346.287400@o13g2000cwo.googlegroups.com>... > > You asked for help and you had someone try to help you in good faith. > > From your response I would never give you any assistance in the future. > > me neither. And 'missing your help' Superboer would be? What? Not missed? [SNIP] > specially since this is Dutch and comes from Amsterdam. [SNIP] And very proud of Both, little boy. [SNIP] > and i have a good memory for these things. > Superboer. [SNIP] Yeah, you would and should, it's a noteworthy experience, meeting me. I really am a very nice guy, said the cat as he approached the mouse, smiling and flashing long white sharp fangs. Robert A. Reissaus IBM Informix Consultants Risdi Amsterdam, Brussel, Chicago, Zurich