Re: GIGE connectivity via TSM
If it the client GigE -- Server IP is not the default route in both directions, and they are not on the same subnet, you will have to configure a static route on each host in order to guarantee that the traffic will traverse the chosen route in both directions. IOW, static routes on both client and server are necessary. -- Tom Thomas A. La Porte, DreamWorks SKG mailto:[EMAIL PROTECTED] On Tue, 23 Mar 2004, Wholey, Joseph (IDS DMDS) wrote: Bill, I hate to beat this... but my situationa is as follows: I have a TSM server residing on Z/OS with an FQDN and one IP address (as far as I know). My client has multiple NICs. We want to use a particular GIGE card exclusively for backup/restore and the others for applications, so we can't disable the others. My server's IP is not on the same subnet as my client's GIGE card, so the default route would not be over the GIGE. QUESTIONS: 1. Do I need to statically route the client GIGE ip address to the Server Ip address in the client's routing table? 2. Does this ensure that data coming from the server to the client will come down that path as well. (I don't want to fall into your I had a client... scenario. thx. joe -Original Message- From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] Behalf Of Bill Boyer Sent: Tuesday, March 23, 2004 2:08 PM To: [EMAIL PROTECTED] Subject: Re: GIGE connectivity via TSM If you use the GIGE address of the TSM server in your DSM.OPT file on the client, then the BACKUP data will go over the GIGE card. Any data FROM the TSM server TO the client will go out the network adapter on the TSM server however TCPIP routing takes it. If you have multiple adapters in the TSM server, the outbound traffic will first go the adapter that is on the same subnet as the client. If the GIGE and client are not on the same subnet, then the traffic will go out the default route unless you have a specific ROUTE in effect. If you want ALL data to go in and out the GIGE card, either disable the other adapters, or change the default route to be the GIGE adapter. I had a client that had in their AIX TSM server 1 10/100 card and 2 GIGE cards ALL set to the same subnet. Backup inbound data to the TSM server came in over the GIGE adapter, but since the 10/100 card was the default route, all outbound traffic, and that included RESTOREs went out the 10/100. They couldn't figure out why restores too so long. Bill Boyer DSS, Inc. -Original Message- From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] Behalf Of Dwight Cook Sent: Tuesday, March 23, 2004 1:18 PM To: [EMAIL PROTECTED] Subject: Re: GIGE connectivity via TSM Use a traceroute command from your client box to the specific ip address you want to go into on your tsm server. Your backups will take the same path. That simple... Dwight E. Cook Systems Management Integration Professional, Advanced Integrated Storage Management TSM Administration (918) 925-8045 Wholey, Joseph (IDS DMDS) To: [EMAIL PROTECTED] [EMAIL PROTECTED]cc: .COMSubject: Re: GIGE connectivity via TSM Sent by: ADSM: Dist Stor Manager [EMAIL PROTECTED] .EDU 03/23/2004 10:28 AM Please respond to ADSM: Dist Stor Manager OK... now things are getting fuzzy. This was my initial concern. Can anyone confirm exactly what one needs to do to ensure one traverses the GIGE NIC for both inbound and outbound traffic? -Original Message- From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] Behalf Of Dwight Cook Sent: Tuesday, March 23, 2004 10:50 AM To: [EMAIL PROTECTED] Subject: Re: GIGE connectivity via TSM tcpclientaddress only sets the initial address for the server to come in on (last I remember) beyond that, standard system/network routing takes over. I'm still at 4.2 5.1 (moving to 5.2 in the next couple of months) but we currently put route statements on the client nodes to ensure they exit out their proper interfaces to access the tsm server(s). Dwight E. Cook Systems Management Integration Professional, Advanced Integrated Storage Management TSM Administration (918) 925-8045 Wholey, Joseph (IDS DMDS) To: [EMAIL PROTECTED] [EMAIL PROTECTED]cc: .COMSubject: GIGE connectivity via TSM Sent by: ADSM: Dist Stor Manager [EMAIL PROTECTED] .EDU 03/23/2004 09:28 AM Please respond
Re: 3494 Library and dual gripper?
The dual-gripper is considered a high availability option, thus each gripper must have access to all cartridges in the library. Thus, the top two rows of each frame are blocked off because the second gripper can not reach them. The bottom two rows are similiarly blocked because the top gripper can not reach them. -- Tom Thomas A. La Porte, DreamWorks SKG mailto:[EMAIL PROTECTED] On Sat, 25 Oct 2003, Hart, Charles wrote: How would you loose slots being that the secongripped will be on the robot ? I'm clearly missing somethig.
Re: New to TSM
Agreed. We have three AIX servers in our studio--out of over 1000 servers and workstations--and those are our TSM servers. Granted, we are a Unix shop, but as most non-AIX Unix admins will tell you, AIX can be a very different beast. Nevertheless, the amount of adminstrative work we have had to do on these boxes over the past five years (yes, *five* years) has been negligible. To bring in another thread, we are moving our TSM servers to Linux, but not as a result of a disappointment with AIX, simply keeping in line with corporate strategy. -- Tom Thomas A. La Porte, Dreamworks SKG mailto:[EMAIL PROTECTED] On Thu, 12 Jun 2003, Dan Goodman wrote: Mark Cini wrote: ... We haven't purchased any hardware yet for TSM, as management wants to be sure we pick the right combination. I am leaning towards TSM server running on Win2000 because of our current in-house expertise. We are excited about the capabilities of TSM combined with an estimated $100k savings over the next year in hardware replacement costs alone. Have you considered that perhaps the reason you have more inhouse Win2k expertise is because Win2k requires more support expertise? Don't forget to factor in the Win2k support costs, and be sure to get actuals from other users, not just published whitepapers. A word to the wise... AIX is a scalable, reliable workhorse, and is the native platform for TSM. These are not things that should not be overlooked, IMHO. Dan Goodman Systems Engineer Specialist Thomas Jefferson University Hospital 215-503-6808 Daniel.Goodman AT mail.tju.edu
Re: Clear text passwords. Was: Automating dsmserv
Since this topic of clear text passwords has arisen, I wonder if anybody knows whether or not there is/are any outstanding requirements or enhancement requests for Kerberos support within TSM. This would be handy both in the situation discussed below, and for general administrative and node access to the server. If there isn't an outstanding request, I'll probably go ahead and ask that one be made. -- Tom Thomas A. La Porte, DreamWorks SKG mailto:[EMAIL PROTECTED] On Tue, 27 May 2003, Richard Sims wrote: Is it just me or does everyone think that placing sensitive userids and passwords in clear text is just a bad thing? It's bad. Not complaining about the procedure here, you gotta do what you gotta do, but has anyone complained to IBM about requiring clear text passwords for this and other scripts? There is no requirement for passwords to be in scripts - that was just someone's conventional implementation of a convenience script. Discretionary halts of the TSM server - as is the case with any daemon style application - are best done via its conventional adminstrative means: in TSM that is disabling sessions thereafter doing a Halt. (Try to avoid thinking that everything running in a Unix environment should be controllable via some rc script.) During a TSM install, that process plants a server start-up method appropriate to the environment, such as /etc/inittab in AIX. One can emulate whatever that is in a superuser invocation to start the TSM server. In traditional Unix, halting the TSM server can be achieved automatically during Unix shutdown via /etc/rc.shutdown , wherein that root-accesssible-only script would contain a dsmadmc command with passsword. It is also conventional in Unix implementations of the TSM server that the server shuts itself down cleanly when it receives a SIGTERM signal (the default signal issued by the Unix 'kill' command). In thinking about sensitive information related to servers in general, consider that, pretty much by definition, a server should be physically secure and not be a system used by ordinary users. Files containing sensitive information should have directory and file permissions which restrict access by those needing it. And where passwords need apply, various means can be employed to avoid having to code them into files (sudo, proxy, etc.). Richard Sims, BU
3494 library on earthquake pads?
I realize that this isn't *strictly* TSM-related, but I know that there are many 3494 library users and some of you may have investigated this before me. We are about to move our 3494 library within our data center and are considering moving it onto seismic earthquake pads (see URL:http://www.worksafetech.com/pages/isobase.html, for an illustration of what I'm describing). I would like to know if anybody on the list has done this before or has investigated doing anything similar. We have been moving all of our data center equipment onto ISO-Base platforms, but the 3494 library seems as though it might have different requirements, what with it's need to stay aligned, etc. Feel free to contact me directly or respond to the list if you have had any experience in this area. -- Tom Thomas A. La Porte, DreamWorks SKG mailto:[EMAIL PROTECTED]
Re: Clear text passwords. Was: Automating dsmserv
On Tue, 27 May 2003, Stapleton, Mark wrote: One of the nice things about how Tivoli has handled TSM is that the authentication system is *exactly* the same, no matter what the server and client OS platforms may be. The same can be said for the interfaces and the way administration is performed. Inserting something like Kerberos into the mix would mean you'd have to make it work for all platforms that the TSM server supports--including MVS, OS/400, and shudder Windows. I'm not suggesting that Kerberos should be required for use in TSM, just that it would be nice if TSM supported it. Having said that, though, we have a mixed Unix/PC/Macintosh environment, and we support Kerberos on all of these platforms. With Win2K, it's essentially built-in to the OS, so I should think that that would be the platform with the least worries. As far as support for other platforms, Kerberos runs on all of the server platforms that the most recent versions of Tivoli Storage Manager supports: Windows NT/2000, AIX, HP-UX, Solaris, MVS/OS390, and Linux. And the good bit is that all of those Kerberos implementations share a common API, so it should not involve much coding to make it work on all of the platforms. There are ways of scripting TSM tasks that can sidestep the clear text stuff, much the same as the ways you script FTP sessions without putting passwords where users can gefingerpoken. True. I was merely suggesting that using Kerberos could solve the problem in a conventional and secure manner. -- Tom Thomas A. La Porte, DreamWorks SKG mailto:[EMAIL PROTECTED]
Re: FastT700 with FlashCopy
On Thu, 27 Mar 2003, Jin Bae Chi wrote:
Suad, Thanks for your help. Let me see if I understand correctly.
1. When you said putting db in backup mode, I can think of
ARCHIVELOG mode for Oracle, which will take a few seconds and
execute the flashcopy;
Q: I heard flashcopy takes a lot less time to create a
snapshot. Any analogy about how fast? How much time would you
gain? Let's say mine takes an hour to hot-backup Oracle through
TDP.
The database should already be running in ARCHIVELOG mode--you
can not make a useful hot/online backup without it. It is also
not something that can be changed while the database is open,
only when it is mounted, so switching between ARCHIVELOG and
NOARCHIVELOG mode would require taking the database down and up.
What Suad meant by putting the db into backup mode was to set
each of the tablespaces into archive mode by issuing the command
'alter tablespace TABLESPACENAME begin backup;'
Presuming that flashcopy is something similar to a Netapp
snapshot, you would put all of your tablespaces in backup mode,
execute the flash copy. Again, presuming it performs similar to a
Netapp snapshot, you should expect that process to be very
short--on the order of seconds to a minute or so depending on the
size of the database. Once the flashcopy is completed, you can
take the tablespaces out of backup mode ('alter tablespace TSPACE
end backup;'). At this point you can simplly transfer the
flashcopy files to TSM via a regular backup interface.
2. Change to NOARCHIVELOG mode for Oracle and mount pseudo volume and backup that
filesystem.
Q: Should I create a separate set of disks for fs (pseudu vol)? Can I do LV backup?
Q: Can I use, after backing it up, the same fs for testing or report purposes before
delete it?
Sorry for too many questions. Thanks again.
Hmmm, not quite sure what you're asking here, but I would
recommend a trip through the Oracle docs regarding standby
databases if what you want is a separate reporting instance.
Re: Fragmented Database Maybe?
On Fri, 28 Feb 2003, Farren Minns wrote: Hi Richard and thanks for the response. [...] However, the reason for my post was more that I was concerned about the number of bytes moved figures (below). The 5000Mb volume is either plainly being shown incorrectly, or perhaps something more serious is afoot. Is this something anyone has seen before? 5000Mb - 947,912,704 2500Mb - 2,621,440,000 1000Mb - 1,048,576,000 What you are seeing is the difference between assigned capacity and percentage utilized (see the output of q db). It would appear that your database has an assigned capacity of 8500MB (5000 + 2500 + 1000), and a percent utilization of 51.8%, with the second and third of your data volumes being full, and all new database pages being allocated from the 5000MB volume. -- Tom Thomas A. La Porte, DreamWorks SKG mailto:[EMAIL PROTECTED]
Kerberos support [was Re: password encryption]
While we're on the subject of passwords and password encryption, is there any chance that TSM might support Kerberos in a future release? -- Tom Thomas A. La Porte, DreamWorks SKG mailto:[EMAIL PROTECTED] On Wed, 19 Feb 2003, Seay, Paul wrote: In encryption speak. The node name is usually called the public key. The private key is what is used to encrypt the message. This is a nice implementation because during password change (which is probably in the message) the new encyption key (password) is not exposed. Paul D. Seay, Jr. Technical Specialist Northrop Grumman Information Technology 757-688-8180 -Original Message- From: Andrew Raibeck [mailto:[EMAIL PROTECTED]] Sent: Wednesday, February 19, 2003 8:02 PM To: [EMAIL PROTECTED] Subject: Re: password encryption To clarify my earlier response on this: The (encrypted) password is not actually sent between client and server, except when the password is being changed. During authentication, the client sends the server a message that is encrypted using the password as the key. The server knows what the decrypted message should be, so if the wrong password was used to encrypt the message, then the authentication will fail. Regards, Andy Andy Raibeck IBM Software Group Tivoli Storage Manager Client Development Internal Notes e-mail: Andrew Raibeck/Tucson/IBM@IBMUS Internet e-mail: [EMAIL PROTECTED] (change eye to i to reply) The only dumb question is the one that goes unasked. The command line is your friend. Good enough is the enemy of excellence. Andrew Raibeck/Tucson/IBM@IBMUS Sent by: ADSM: Dist Stor Manager [EMAIL PROTECTED] 02/19/2003 14:56 Please respond to ADSM: Dist Stor Manager To: [EMAIL PROTECTED] cc: Subject:Re: password encryption The password is indeed encrypted. Regards, Andy Andy Raibeck IBM Software Group Tivoli Storage Manager Client Development Internal Notes e-mail: Andrew Raibeck/Tucson/IBM@IBMUS Internet e-mail: [EMAIL PROTECTED] (change eye to i to reply) The only dumb question is the one that goes unasked. The command line is your friend. Good enough is the enemy of excellence. Prather, Wanda [EMAIL PROTECTED] Sent by: ADSM: Dist Stor Manager [EMAIL PROTECTED] 02/19/2003 14:40 Please respond to ADSM: Dist Stor Manager To: [EMAIL PROTECTED] cc: Subject:Re: password encryption I've always been told that the password is NOT sent in plain text, it's encrypted. (but I've never had a sniffer to check it myself). -Original Message- From: Eliza Lau [mailto:[EMAIL PROTECTED]] Sent: Wednesday, February 19, 2003 10:36 AM To: [EMAIL PROTECTED] Subject: password encryption Does anyone know how the stored password on the client machine is passed to the server for authentication? The user has 'password generate' in his dsm.opt. The password is stored in the Registry of his Windows 2000 client. When the TSM client starts is the password sent to the server in plain text or encrypted? Thanks, Eliza Lau Virginia Tech Computing Center 1700 Pratt Drive Blacksburg, VA 24060
AUTOMOUNT options file option?
Has anybody successfully used the AUTOMOUNT option in a client options file, along with the 'DOMAIN all-auto-nfs' statement? Regardless of what I choose in my AUTOMOUNT directive, I receive the following error message: 09/25/02 16:55:21 Filesystem /studio/sin/mstr/general is no automounted filesystem, ignoring option automount for this filesystem. As can be seen from the mnttab file, this is indeed an automounted filesystem: # grep /studio/sin/mstr /etc/mnttab auto.studio-sin-mstr/studio/sin/mstrautofs indirect,ignore,dev=428012b 1032470327 I get the same behavior on an TSM 4.2 client on Irix, TSM 5.1 on Solaris, and TSM 4.2 and 5.2 on Linux. I have also tried it for my home directory 'AUTOMOUNT /usr/home/tlaporte' with similar results. For what, exactly, is the AUTOMOUNT directive looking? The frustrating point is that 'dsmc incr -domain=/usr/home/tlaporte' seems to work fine, and a query of that node reports a filespace type of NFS3. Thanks for anybody's thoughts or experiences on the matter. -- Tom Thomas A. La Porte, DreamWorks SKG mailto:[EMAIL PROTECTED]
Re: AUTOMOUNT options file option?
On Wed, 25 Sep 2002, Chetan H. Ravnikar wrote: tell me what you want to achive and I will try and help you. I had the same probelm to get autmount working on solaris-2.8 and TSM 4.2.20. Infact never got it to work., But I have a workarround. In my environment, I have a design to backup auto-mounted netapp filers snapshot directories over NFS Cheers.. We are also backing up Netapp filers over NFS, and we have used several solutions, none of which are completely satisfactory to us. We are trying to see if we can't work under a supported configuration, since TSM claims to support automounted filesystems..
Re: Eternal Data retention brainstorming.....
As a refinement to Wanda's final suggestion, couldn't you alter your policies for 'del volhist type=dbb' (or simply retain the current copy of your database backup exclusive of the volume history), and then modify your storage pool's reusedelay parameter appropriately? The drawback that I see is that there is no forever parameter for reusedelay. days is the maximum. Granted, that's over 27 years, but we know how long government investigations can last! :-) -- Tom Thomas A. La Porte, DreamWorks SKG mailto:[EMAIL PROTECTED] On Thu, 15 Aug 2002, Prather, Wanda wrote: Besides that, the best solution I can think of, change all the management classes to never expire/unlimited versions, Copy the DB to your test server, lock all the client nodes, put your tapes on a shelf. Save the last DB backup, just in case. Start your production server over with a clean DB, back up everything new and move on. If anybody needs their old stuff, get a copy via export (from test) and import(back to production). That would keep you from (immediately) doubling your tape requirements, will cost you some hardware to make tape available for your test system.. Wanda Prather
Re: Help with select statement
I'll take a stab: select cast(entity as varchar(12)) as Node Name, \ cast(activity as varchar(10)) as Type, \ sum(cast(affected as decimal(7,0))) as files, \ sum (cast(bytes/1024/1024 as decimal(12,4))) as Phy_MB \ from summary where end_time=timestamp(current_date -1 days, '09:00:00') \ and end_time=timestamp(current_date, '08:59:59') \ and (activity='BACKUP' or activity='ARCHIVE') \ group by entity, activity \ order by Node Name On Thu, 18 Jul 2002, L'Huillier, Denis wrote: Hello - I wrote the following select statement (with a lot of plagiarism). /* --- Query Summary Table */ select cast(entity as varchar(12)) as Node Name, \ cast(activity as varchar(10)) as Type, \ cast(affected as decimal(7,0)) as files, \ cast(bytes/1024/1024 as decimal(12,4)) as Phy_MB \ from summary where end_time=timestamp(current_date -1 days, '09:00:00') \ and end_time=timestamp(current_date, '08:59:59') \ and (activity='BACKUP' or activity='ARCHIVE') \ order by Node Name The problem is, if a node performed 10 backups and 5 archives over the 24 hour period there are 15 lines for that node in the output, 10 for backup and 5 for archive. Is there a way I can add the affected columns and bytes column for a node which has activity=BACKUP and again for those with activity=ARCHIVE ? Basically, what I want is at most 2 lines per node... 1 line can be the sum of affected files and bytes for all backup activities And the other line for that node can be the sum of affected files and bytes for all ARCHIVE activities. I think I'm over my head. Regards, Denis L. L'Huillier 212-647-2168
Linux IA-64 client?
Anybody know of any plans for a TSM client for Linux on the IA-64 platform? Thanks. Thomas A. La Porte, DreamWorks SKG mailto:[EMAIL PROTECTED]
Re: Ordering license upgrades
And while we're on the subject of licenses and upgrades, I was informed by my reseller that Tivoli has withdrawn the TSM Enterprise Edition from marketing. All of the constituent parts are still available, but you can not purchase the bundle. Word is that it was too confusing to end users. I hope it didn't also have to do with the fact that our software quote went up by $70,000 when the change was made. -- Tom Thomas A. La Porte, DreamWorks SKG mailto:[EMAIL PROTECTED] On Sat, 18 May 2002, Zlatko Krastev wrote: 'cause there is no published rule how to convert points to processors. I am trying to find the answer since the announcement of new licensing scheme but still without success. Just ask them for NN processors or MM points. In both cases you have the right to use them on different platforms. Zlatko Krastev IT Consultant Please respond to ADSM: Dist Stor Manager [EMAIL PROTECTED] Sent by:ADSM: Dist Stor Manager [EMAIL PROTECTED] To: [EMAIL PROTECTED] cc: Subject:Ordering license upgrades We have a TSM 4.2.1.9 server running under OS/390. We are about to order more client licenses. The last time we did this IBM and/or Tivoli required us to tell them how the new licenses would be divided among different platforms: how many for Windows NT, how many for AIX, and so on. The charge per client license is the same regardless of operating system. The 'register license' command does not provide any means for distinguishing between Windows NT licenses, AIX licenses, and so on. Why in the world do we need to provide platform information in order to get more client licenses?
Re: Recovery Log almost 100%
I wonder, also, if there is still any discussion about supporting the use of an alternate RDBMS underneat TSM. It is quite clear that there are many more sites with database sizes in the 25-50GB+ range. Five years ago I felt very lonely with a database of this size, but given the discussions on the listserv over the past year I feel more comfortable that we are no longer one of the only sites supporting TSM instances that large. It has always seemed to me that the database functions of TSM have been the most problematic (deadlock issues, log full issues, SQL query performance problems, complicated and unclear recommendations for physical database layout, etc.). All of these problems have been solved by Oracle, DB2, and Sybase. Granted there is the issue that plugging in an external database adds greatly to the complexity of TSM, and reduces it's black box-ness, but I think the resources are available to administer such a beast at the large sites that require very large databases. More food for thought *early* on a Thursday morning. -- Tom Thomas A. La Porte DreamWorks SKG [EMAIL PROTECTED] On Thu, 2 May 2002, Paul Zarnowski wrote: TSM Development is fully aware of the log issue and based on some conversations at SHARE, I am comfortable that they are taking steps to address it (with or without a requirement). I don't think this issue will be completely solved quickly, as it is a rather complex set of problems. In the short term, look for tools to show up that will help TSM administrators to identify which session has the log tail pinned, and also address one of the issues that Paul refers to below, which causes the log head to advance quickly (and shows up as a high dirty page count). When the log fills, two things happen: The log tail must be pinned by a long-running in-flight transaction, and the log head must advance around to catch up to the tail. To keep the log from filling, you can either release the tail or slow down the head. It is not easy to identify the session or thread that has the log tail pinned. I don't know if the tools I refer to above have shown up in 4.2.2 or 5.1 (we're still running 4.2.1). There are a couple of things that can advance the head quickly. Inventory expiration and filespace deletion. If you find yourself in a situation where you see the log filling quickly and don't know what has the tail pinned, check for these two processes and kill them if you see them. This will significantly slow down the growth rate of the log, and give the oldest in-flight transaction more of a chance to complete. We have written a monitor to do this automatically, and it has really helped us. If neither of these processes are running, then you can start guessing about which session might have the tail pinned. In this situation, we look for an old session that has been running for a long time. This might be a session backing up over a slow speed line. If the log nears 100%, we try to avoid it filling completely by cancelling all sessions (if we have time) or simply HALTing the server and restarting it. This generally clears the log when the server comes back up, and avoids having to do an offline extend of the log (which has already been discussed). If you are running logmode=rollforward, be aware that when you later reduce the log size to delete the temporary extension, you will (I think) trigger a full database backup. If you are at v4.2, you can have a larger log, up to 13GB. This can also provide some relief. ..Paul At 12:13 AM 5/2/2002 -0400, Seay, Paul wrote: Actually, this was significantly discussed at Share and the basic requirement is TSM, take action whatever necessary to keep the server up. Start by cancelling expiration. Then nail the client that has the log pinned. There were also a number of issues discussed. Apparently, there are a lot of dirty blocks being recorded in the log that do not have to be. I am working to get these requirements voted on. Paul D. Seay, Jr. Technical Specialist Naptheon, INC 757-688-8180 -Original Message- From: Thomas A. La Porte [mailto:[EMAIL PROTECTED]] Sent: Wednesday, May 01, 2002 4:36 PM To: [EMAIL PROTECTED] Subject: Re: Recovery Log almost 100% Given that this is one of the more comman FAQ style questions on this listserv, I wonder if it's not time for someone to submit a TSM requirement that the server behave better in a recovery log full situation. This happens in other databases w/o causing a SIGSEGV. Oracle, for example, simply prevents any database changes, and only allows new administrative connections to the database until the log full situation is cleared (by archiving the online redo logs). It seems that TSM could behave similarly. Certainly the server is not in a great state when the log segments are full, but it would seem easier to recover, and somewhat less confusing to administrators, if it could be done online, rather than in the manner in which it is handled now. We've all probably
Re: Recovery Log almost 100%
Given that this is one of the more comman FAQ style questions on this listserv, I wonder if it's not time for someone to submit a TSM requirement that the server behave better in a recovery log full situation. This happens in other databases w/o causing a SIGSEGV. Oracle, for example, simply prevents any database changes, and only allows new administrative connections to the database until the log full situation is cleared (by archiving the online redo logs). It seems that TSM could behave similarly. Certainly the server is not in a great state when the log segments are full, but it would seem easier to recover, and somewhat less confusing to administrators, if it could be done online, rather than in the manner in which it is handled now. We've all probably experienced a situation where we are close to the limit on the log size, so we only extend the log a little bit, and then there is a rush to see if our database backup is going to finish and clear the log full condition before we use up the additional log space--lest we find ourselves in the same perilous condition, only *closer* to the seemingly arbitrary maximum log size. -- Tom Thomas A. La Porte DreamWorks SKG [EMAIL PROTECTED] On Wed, 1 May 2002, Sung Y Lee wrote: When log reaches 100%, just pray that TSM server process will not crash. I say the key is prevention. Whatever you can do to prevent that from happening is the best answer. There are many things you can do to prevent from growing to 100%. One that works for me is I have LogMode set to Roll Forward mode with dbb trigger at 38% with incremental between at 3(q dbb) Log is also set to maximum allowed without going over limit plus room for extension should it ever reaches 100% and TSM crashes. Have it set at 4.5 GB(To be safe). Max allowed recovery log for TSM 4.1 is 5.3 GB?? I can't recall exact value. If the TSM server is in Log mode than more than likely it will have dbb trigger set at certain point. For example, adsm q dbb Full Incremental Log Full Incrementals Device Device Percentage Between Class ClassFulls -- --- -- IBM3590IBM3590 38 3 When the recovery log reaches 38%, an incremental database backup will kick off up to 3 threes before a full database backup performed. The most of time TSM server will prevent other sessions from establishing when the recovery log reaches 100% but will allow the trigger database backup to complete and it will bring the recovery log down to 0. Sometimes TSM will simply crash. If it crashes then you will need to do an emergency recovery log extend and bring TSM backup. Sung Y. Lee E-mail [EMAIL PROTECTED] brian welsh brianwelsh3@HOTMTo: [EMAIL PROTECTED] AIL.COM cc: Sent by: ADSM: Subject: Recovery Log almost 100% Dist Stor Manager [EMAIL PROTECTED] .EDU 05/01/2002 01:23 PM Please respond to ADSM: Dist Stor Manager Hello, AIX 4.3.3 and server 4.1.1.0. Last night two archive-schedules had a problem. On the clients there were files in a kind of loop and TSM tried to archive them. Result, recovery log almost 100%. This was the first time our log is that high. Problem on the client solved, but now I have the following question. I was wondering how other people prevent the log from growing to 100%, and how to handle after the log have reached 100%. Any tip is welcome. Brian. _ MSN Foto's is de makkelijkste manier om je foto's te delen met anderen en af te drukken: http://photos.msn.nl/Support/WorldWide.aspx
Re: About TSM API
Fred, If by talking to the server without a client you mean running administrative commands, there is no TSM administrative API. The client API is for backup/restore and archive/retrieve operations, and that is the only API available. -- Tom Thomas A. La Porte DreamWorks SKG [EMAIL PROTECTED] On Mon, 29 Apr 2002, Fred Zhang wrote: Hi, I am looking for TSM API. So far I can only find Client API. Does TSM has API which we can use to talk to the server without a client installed? Any information would be greatly appreciated. = Fred Zhang NetiQ Corporation 3553 N. First St. San Jose, CA 95134 phone: (408)856-3102 fax: (408)856-3102 e-mail: [EMAIL PROTECTED] =
Re: ANR0534W - size estimate exceeded
We have seen this behavior in an API client that we have written, and we have verified that (a) COMPRESSION is set to NO and (b) we are sending exactly the same number of bytes that we say we are sending. We have determined that there is essentially a race condition when there are multiple sessions writing to the disk pool. If, at a given point in time, four clients contact the server and announce the intention to write a 100MB, the server checks at the beginning of each of the four sendObj() calls. If there is 150MB of space available in the diskpool at the time that each sendObj() call happens (API programmers forgive me if I've gotten the calls slightly incorrect, I've not written the API client, only assisted as the TSM admin), the server gives each of the four clients the OK to send the data. Of course, they will all likely fail, as they will fill up the 150MB availables space before any *one* of the clients finishes. This has been one of the frustrating aspects of the TSM BA client not being written with the TSM API. Clearly the BA client does not encounter this problem, yet there is no straightforward method to solve the problem with the API. I believe that the TDP products are all written with the API, so I am not surprised to hear that they suffer from this problem, as well. AFAIK, the best we have been able to get from support is the admission that the error message is inaccurate, or at the very least misleading. -- Tom Thomas A. La Porte DreamWorks SKG [EMAIL PROTECTED] On Fri, 26 Apr 2002, Tomas Hrouda wrote: Gerhardt, sending of incorrect filesize obviously happens (by my experiences) when commpression is turned on and compressed file grows up to original file. Try to use no compression or COMPRESSALWAYS NO to avoid bad file size prediction by client. It was discussed here sometimes in the past. Hope this helps. Tom -Original Message- From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED]]On Behalf Of Gerhard Wolkerstorfer Sent: Friday, April 26, 2002 12:39 PM To: [EMAIL PROTECTED] Subject: Antwort: Re: Antwort: ANR0534W - size estimate exceeded Zlatko, you are right, BUT..when the TDP sends an incorrect Filesize to the TSM Server, the maxsize Parameter won't work (TDP sends 100 Byte - the server will let the File go to the diskpool, but the file will indeed have 20 Gb, which will fill up your diskpool and bring up the message indicated (storage exceeded)) And for tracing purposes I wanted to know, if there is any possibility to check the filesize, which the client is sending to the server. Gerhard Wolkerstorfer [EMAIL PROTECTED] (Zlatko Krastev) am 26.04.2002 11:34:22 Bitte antworten an [EMAIL PROTECTED] An: [EMAIL PROTECTED] Kopie: (Blindkopie: Gerhard Wolkerstorfer/DEBIS/EDVG/AT) Thema:Re: Antwort: ANR0534W - size estimate exceeded Isabel, Gerhard, you can set MAXSIze parameter of the disk pool. I usually set it about 30-60% of the diskpool size (or better pool free size, i.e. size - highmig). Files larger than this would bypass the diskpool and go down the hierarchy (next stgpool). For this might be tape pool, i.e. file will go direct to tape. Zlatko Krastev IT Consultant Please respond to ADSM: Dist Stor Manager [EMAIL PROTECTED] Sent by:ADSM: Dist Stor Manager [EMAIL PROTECTED] To: [EMAIL PROTECTED] cc: Subject:Antwort: ANR0534W - size estimate exceeded Isabel, we still have this problem regarding to the TDP Informix. It seems, that the TDP (sometimes ?) isn't sending the correct Filesize and the File (DB Backup) exceeds your DISKPOOL and cannot swap to the Tapepool. If the TDP would send the correct Filesize, TSM would possibly go direct to tape and the problem wouldn't come. Question: How can I check the Filesize and/or Filename, the client is sending to the server ?? MfG Gerhard Wolkerstorfer [EMAIL PROTECTED] (Isabel Gerhardt) am 26.04.2002 09:17:44 Bitte antworten an [EMAIL PROTECTED] An: [EMAIL PROTECTED] Kopie: (Blindkopie: Gerhard Wolkerstorfer/DEBIS/EDVG/AT) Thema:ANR0534W - size estimate exceeded Hi Listmembers, we recently started to recieve following errors: 04/23/02 20:30:29 ANR0534W Transaction failed for session 1 for node NODE1 (WinNT) - size estimate exceeded and server is unable to obtain additional space in storage pool DISKPOOL. 04/24/02 20:38:19 ANR0534W Transaction failed for session 173 for node NODE2 (TDP Infmx AIX42) - size estimate exceeded and server is unable to obtain additional space in storage pool DISKPOOL. From previous Messeges of the List I checked, that the Diskpool has caching disabled and the clients have no compression allowed. I was away from work for a while and meanwhile a serverupdate has been done. If anyone can point me to the source of this error, please help! Thanks in advance, Isabel
Re: TDP for NDMP
We are definitely anxious to see file-level granularity. This (NDMP support) is something that Legato has had for well over a year now, so TSM is a bit behind in its support. Given the success of Network Appliance, I would think that there is becoming a large enough user base that is interested in this support. -- Tom Thomas A. La Porte DreamWorks SKG [EMAIL PROTECTED] On Tue, 23 Apr 2002, Don France (TSMnews) wrote: You've just about gottit all. There is some concern about this whole thing; currently, you are limited to image backups using NDMP; if you want file-level granularity, you must use NFS mount to access the file system from a supported b/a-client... not as pretty as you'd like, but that's the way it is. Then, too, you are limited to NAS filer; Auspex and other filers are not supported. The storage pool limitation comes from the format used to store the data; they invented a new concept/format for NDMP storage pools, so must be kept segregated from standard storage pools; I THINK you can still use backup stg command -- but you cannot intermix the different types of primary storage pools in a common copy-pool. IBM/Tivoli is wondering how much market there is to justify expanding this support -- (a) file-level granularity and/or (b) other filers. IMHO, I'd want to see file-level granularity. Comments can be posted here... Don France Technical Architect - Tivoli Certified Consultant Professional Association of Contract Employees (P.A.C.E.) San Jose, CA (408) 257-3037 [EMAIL PROTECTED] - Original Message - From: Gerald Wichmann [EMAIL PROTECTED] To: [EMAIL PROTECTED] Sent: Wednesday, April 17, 2002 9:47 AM Subject: TDP for NDMP Has anyone used and had any experiences with NDMP backups? I'm curious what kind of experiences you've had using the product. Pros/cons, etc.. I've been reading the admin guide section on it and it seems like a whole other beast to what I'm used to with TSM. E.g. your NDMP file server has to be attacked to the tape drive. Seems like this would be undesirable if you had a lot of file servers. Would you have to buy a comparative # of tape drives? Or how many tape drives could you share between your file servers and TSM server? Theres a comment about you cannot backup a storage pool used for NDMP backups. Does this mean you can't create a copypool and 2nd copy of the backed up data? Do the NDMP backups only backup the entire file server and restore the entire file server? I guess it's not clear to me on what level of granularity there is in backing up NDMP file servers. Seems like NDMP backups require their own TSM domain/policyset/etc structure as well. Also I understand it only works with Network Appliance NAS devices? Appreciate anyone with practical knowledge to comment on their experience and perhaps clarify some of my questions above.. thank-you
Re: How Best to Suspend a Schedule
That would involve recycling the server, wouldn't it? I'm guessing from Paul's e-mail that he's looking for something he can do online. What I've done in the past is update the start date of the schedule to the following day, that way it won't run in the coming evening, but will pick up again the next day without me having to intervene. On Fri, 19 Apr 2002, Edgardo Moso wrote: For #1 you can set , disablesched Yes in the dsmserv.opt file. From: Kilmer, Paul [EMAIL PROTECTED] on 04/19/2002 09:02 AM Please respond to ADSM: Dist Stor Manager [EMAIL PROTECTED] To: [EMAIL PROTECTED] cc: Subject: How Best to Suspend a Schedule As a relative newbie to TSM, I'd be interested to hear what is the easiest way to temporarily: 1- Suspend a backup schedule from running. 2- Suspend an individual client's participation in a backup schedule. I've accomplished #2 by removing the client schedule association, though I don't regard this method to be particularly elegant. Thanks in advance. Paul E. Kilmer Lead Technical Architect New World Pasta [EMAIL PROTECTED] (717) 526-2431 || |*http://www.eteamz.com/cdboys www.eteamz.com/cdboys *| || ||
Re: Help Urgent!
This often means that the tape volume on which the object resides in in status UNAVAILABLE. Check your activity log on the server at the time of the restore. You should see additional information. -- Tom Thomas A. La Porte DreamWorks SKG [EMAIL PROTECTED] On Thu, 11 Apr 2002, Amini, Mehdi wrote: While restoring Win2K systemobject, I got messages Data unavailable to Server What does this mean? Thanks Mehdi Amini LAN/WAN Engineer ValueOptions 3110 Fairview Park Drive Falls Church, VA 22042 Phone: 703-208-8754 Fax: 703-205-6879 ** This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error please notify the sender by email, delete and destroy this message and its attachments. **
Re: Restore TSM server to different OS?
Search the archives. Although it may be technically possible to do this move (though I doubt that you could go from Windows to a Unix platform), it is definitely not supported. Wanda Prather had an excellent summary of the discussion when this came up again last fall: http://msgs.adsm.org/cgi-bin/get/adsm0111/951.html -- Tom Thomas A. La Porte DreamWorks SKG [EMAIL PROTECTED] On Tue, 19 Mar 2002, Daniel Sparrman wrote: Hi Restoring the TSM server(database) to a different OS will work fine as long as you keep to an open systems platform such as Windows NT/2000, HP-UX, AIX or Solaris. Doing this on a AS/400 or S/390 wont work. This will apply to doing a restore of the TSM database. Best Regards Daniel Sparrman --- Daniel Sparrman Exist i Stockholm AB Bergkällavägen 31D 192 79 SOLLENTUNA Växel: 08 - 754 98 00 Mobil: 070 - 399 27 51 -ADSM: Dist Stor Manager [EMAIL PROTECTED] wrote: - To: [EMAIL PROTECTED] From: Hans Hedlund [EMAIL PROTECTED] Sent by: ADSM: Dist Stor Manager [EMAIL PROTECTED] Date: 03/19/2002 06:15PM Subject: Restore TSM server to different OS? Is it possible to restore a TSM 4.1 server running on Windows NT to a Solaris server? I'm practicing some DR stuff, and my scenario is that one of our WinNT TSM4.1 servers has crashed, and we need to restore the latest BACKUP DB data, but onto a Solaris server on our disastersite. Is that possible? I receive mysterious ANR1368W Input volume 16440215.DBB contains sequence number 16777216; volume sequence number 1 is required. when doing my DSMSERV RESTORE DB. Regards, Hans Hedlund Confidentiality Notice The content of this e-mail is intended only for the confidential use of the person(s) to whom it is addressed. If the reader of this message is not such a person, you are hereby notified that you have received this communication in error and that reading it, copying it, or in any way disseminating its content to any other person, is strictly prohibited. If you have received this message in error, please notify the author by replying to his e-mail immediately.
Re: Oracle on Linux backups
There was an old Redbook on using ADSM to backup databases. It includes some very good examples of using shell scripts to accomplish hot backups of Oracle (and many other databases) without the need for a TDP or RMAN style backup. We use a method like this to backup all of our Oracle instances (Linux or otherwise). None of our instances are large enough to warrant the additional complication of RMAN or TDP. In addition to the Redbook above, if you are at all involved in the backup and restoration of Oracle databases, I would *highly* recommend the Oracle Press book Oracle Backup and Recovery by Velpuri, et al (it exists in several editions for the various versions of Oracle). -- Tom Thomas A. La Porte DreamWorks SKG [EMAIL PROTECTED] On Wed, 13 Mar 2002, Zoltan Forray/AC/VCU wrote: I had earlier asked the question about using the TSM TDP for Oracle on a Linux box but never got a response. From the silence and other messages related to this topic, I can only surmize that this won't/doesn't work and I shouldn't waste my time/money in purchasing the TDP. So, since I can't change the platform, I need to know how anyone else in this situation/configuration backs up their Oracle systems on Linux ? Or are most folks waiting for TSM 5.x and the supposed forthcoming TDP that will work on Linux and are keeping Oracle on supported/backup-able OS'es ? Note, I am not a DB person. I don't know the first thing about Oracle. I have seen some discussions about an RMAN program/utility, but I only know the name. The objective is to do live (i.e. not shutting down the application) backups of the Oracle DB and be able to restore it on a backup/test machine. We do a nightly backup but all attempts to restore from these backups to another box, have produced useless files. Oracle is very unhappy !! Any and all help would be greatly appreciated. Zoltan Forray Virginia Commonwealth University - University Computing Center e-mail: [EMAIL PROTECTED] - voice: 804-828-4807
Re: Backup to multiple tape simultaneously
dsmadmc -id=admin -pa=admin upd stg POOLNAME collocate=filespace On Wed, 13 Mar 2002, ARhoads wrote: except for all of the contention for tapes... - Original Message - From: Bill Smoldt [EMAIL PROTECTED] To: [EMAIL PROTECTED] Sent: Wednesday, March 13, 2002 10:08 AM Subject: Re: Backup to multiple tape simultaneously Although it is very easy to start a dsmc restore filesystem for each filesystem and do parallel restores today. Bill Smoldt STORServer, Inc.
API / Client compatibility
We are in the process of writing an Linux archive client through
the TSM API (4.2), storing objects in an TSM 4.1 server running
on AIX. We would like to avoid writing a retrieve client, as
we're satisfied with the functionality of the TSM BA client. I
know there have been interoperability issues in the past, but the
API manual clearly addresses *some* of these issues.
What we are wondering is how file permissions need to be stored
by the API client so that the BA client can retrieve them
properly.
We attempted to pass a struct stat in the objAttrPtr.objInfo
(see below), but when we retrieve the files using the BA client,
all files are retrieved with perms set to 000 and a timestamp
that is very wrong (e.g. Dec 5, 1930 on some files :-), rather
than the perms and timestamp stored in the struct stat.
This was just a guess (which was apparently wrong), but it's
completely unclear as to what *would* be correct. Page 74 of the
API Manual states the following in reference to using
restore/retrieve commands for API objects:
Note: Use these commands for exception situations only.
These commands return data as bit files that are created
using default file attributes. You can restore or retrieve API
objects that other users own, or that are from a different
node. The set access command determines which objects qualify.
If it were truly using default file attributes, I would have
expected them to be umask based. But the fact that perms are set
to 0 and the timestamps are set consistently wrong indicates that
the BA client *is* attempting to use the struct stat info in
objInfo.
Has anybody reconciled this? Has anybody successfully written an
API client for which objects can be retrieved or restored using
the BA client?
Thanks,
-- Tom
Thomas A. La Porte
DreamWorks SKG
[EMAIL PROTECTED]
if ((rc = lstat(filelist[i].filename, sb)) != 0)
{
fprintf(stderr, stat %s: , filelist[i].filename);
perror();
continue;
}
[...]
ObjAttrData.objInfoLength = sizeof(sb);
ObjAttrData.objInfo = (char *)sb
Re: Network Appliance - Filer backup - (DAR) Direct Access Restore
My understanding is that NDMP support will be added to the AIX server in TSM version 5. Don't know if that version changes the requirements of direct-attached. Why Tivoli chose to start with the Windows platform for an enterprise-level protocol like NDMP is still beyond me. We are eagerly anticipating support for NDMP on AIX. -- Tom Thomas A. La Porte DreamWorks SKG [EMAIL PROTECTED] On Wed, 6 Feb 2002, Prather, Wanda wrote: but, from what I have read on the Tivoli web site, you can only get NDMP support if your tsm SERVER is on Windows, and if your tape drives are direct attached to the NAS device, which means you have to be SAN. Am I reading that correctly, or am I missing something here? -Original Message- From: Joshua S. Bassi [mailto:[EMAIL PROTECTED]] Sent: Tuesday, February 05, 2002 7:49 PM To: [EMAIL PROTECTED] Subject: Re: Network Appliance - Filer backup - (DAR) Direct Access Restore TSM support NDMP at this time. -- Joshua S. Bassi Sr. Solutions Architect IBM Certified - AIX/HACMP, SAN, Shark Tivoli Certified Consultant- ADSM/TSM Cell (415) 215-0326 [EMAIL PROTECTED] -Original Message- From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED]] On Behalf Of Gates, Gerald P. Sent: Tuesday, February 05, 2002 4:44 PM To: [EMAIL PROTECTED] Subject: Network Appliance - Filer backup - (DAR) Direct Access Restore Can Tivoli backup up a Network Appliance - Filer using (DAR) Direct Access Restore? Veritas said they will have this in version 5.0 later this year. Gerry
Re: FILESPACE_NAME vs. HL_NAME
Are these by any chance NFS mounted directories? We do a lot of
archives with loopback-mounted filesystems, and over the course
of multiple iterations of the dsmc client we have found that
ADSM/TSM have treated these filesystems differently wrt to what
is FILESPACE_NAME and what is HL_NAME.
The proper way to find these files with a 'dsmc q arch' or
'dsmc retr' is to explicitly name the filespace in your filespec.
So, in your example, you have a single file which has been
archived under three different filespace names. Assuming you were
looking to retrieve the version of the file archived on
2001-12-22, you would issue the following command:
dsmc retr '{/WORK/DATA}/MARTINE/BS0266'
If you do not specify the filespace in your filespec, ADSM/TSM
will use the filespace that has the *longest* match, e.g., in
your case by default it will always find the version archived on
2001-11-20.
-- Tom
Thomas A. La Porte
DreamWorks SKG
[EMAIL PROTECTED]
On 01/02/2002 Mike Crawford wrote:
Good Afternoon,
A client is having trouble locating archived files using dsmc q ar. The
problem
seems to be that the filespace_name and hl_names differ between the files,
even
though they were archived from the same place.
Server: AIX, ADSM v3.1
Client: SunOS, ADSM v3.1
An example:
select filespace_name,hl_name,ll_name,archive_date from archives where
node_name='HUBER'
and ll_name='BS0266'
FILESPACE_NAME: /WORK
HL_NAME: /DATA/MARTINE/
LL_NAME: BS0266
ARCHIVE_DATE: 2001-12-22 10:46:30.00
FILESPACE_NAME: /WORK/DATA
HL_NAME: /MARTINE/
LL_NAME: BS0266
ARCHIVE_DATE: 2001-12-22 10:41:24.00
FILESPACE_NAME: /WORK/DATA/MARTINE
HL_NAME: /
LL_NAME: BS0266
ARCHIVE_DATE: 2001-11-20 05:38:10.00
Depending on how the client specifies the request, they will get a
different
version of the file.
The question is, how does ADSM determine which part of the path is
considered
filespace_name, and which is hl_name?
Thanks,
Mike
Re: Clarifications: LAN Free ( SAN ) Backup
I take it to mean that any given FILE on the SAN device can be written to by only one client at a time, but that many files (up to the MOUNTLIMIT parameter) can be written to at once, meaning the maximum number of clients writing simultaenously to the SAN device is equal to the MOUNTLIMIT parameter in your FILE device class. -- Tom Thomas A. La Porte DreamWorks SKG [EMAIL PROTECTED] On Thu, 15 Nov 2001, Wouter V wrote: So : A FILE devclass device may be in use by only one party., means only one client a time can backup to the SAN-disk device ? The SAN disk is shared, is available for every client, but can not be used concurrent by every client ? Wright ? Wouter.
Re: Merging two server into one
You will require additional storage resources, either tape or disk. The export facility will write all of your TSM data to new storage, and the import will then transfer that data to still new storage (which could be the original storage on the exported server). In any case, you will need to have double the amount of space available for the actual export, because the data will occupy at least two locations for the duration of the export/import process. -- Tom Thomas A. La Porte DreamWorks SKG [EMAIL PROTECTED] On Wed, 31 Oct 2001, Jolley, Bill wrote: If the Private and Scratch Categories are different for each server, how then would the import/export facility handle those? We took over ADSM with two servers located on one AIX SP node and would like to consolidate them. I have sought assistance from TSM support and received very little. Could this be a difficult task? -Original Message- From: Joshua S. Bassi [mailto:[EMAIL PROTECTED]] Sent: Wednesday, October 31, 2001 11:59 AM To: [EMAIL PROTECTED] Subject: Re: Merging two server into one You can use the TSM import/export facility to do this. -- Joshua S. Bassi Independent IT Consultant IBM Certified - AIX/HACMP, SAN, Shark Tivoli Certified Consultant- ADSM/TSM Cell (408)(831) 332-4006 [EMAIL PROTECTED] -Original Message- From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED]] On Behalf Of Andreas Rensch Sent: Wednesday, October 31, 2001 7:54 AM To: [EMAIL PROTECTED] Subject: Merging two server into one Hi TSM-/ADSM-ers, is there a possibility to merge two TSM servers into one? Our situation describes as follows : We have two TSM servers - both at OS/390 - which share the same range of tapes. Now we want to move to the Sun Solaris platform. We planned to take our first TSM server as a base for the new TSM server on Solaris. But we have about 150 tapes on our second TSM server and want to avoid to copy all these tapes. Is there another way to get this data into the first TSM server? One option is to define the clients of the second TSM server at the first TSM server, make first a full backup and then incremental backups. The problem with this option is that we loose our backup history. Any other idea? mfg / regards andreas rensch / StorageManagement / rz-qs phone : +49(0)6171.66.3692 / fax : +49(0)6171.66.7500.3692 / mobile : +49(0)172.6649.016 / pse : 75 / mailto:[EMAIL PROTECTED] Alte Leipziger Lebensversicherung aG - Alte Leipziger Platz 1 - D 61440 Oberursel - http://www.alte-leipziger.de When all else fails, read the instructions.
Re: oracle backup(linux)
Well, like I said, I spoke with my IBM rep and was told that it *could* be available with TSM v5 in the first half of 2002. Since there has been no announcement, we are left to our own design in making determinations about IBM/Tivoli's product directions. It seems safe to assume that a TDPO for Linux might become available, given IBM's heavy push in the Linux arena, but without any announcements I, and in turn my rep, have to seek out this information, and report it back whatever unofficial information that we have garnered. -- Tom Thomas A. La Porte DreamWorks SKG [EMAIL PROTECTED] On Thu, 25 Oct 2001, Thiha Than wrote: hi, I am not sure where you got the information about TDPO for Linux availability in first half of 2002. We are planning for it but don't count on having it in the first half of 2002. regards, Thiha I just recently spoke with my IBM rep on this, and the word I got back that, although there are no specific times set or announced, a TDP for Oracle could be available on Linux with TSM v5 in the first half of 2002.
Re: oracle backup(linux)
I just recently spoke with my IBM rep on this, and the word I got back that, although there are no specific times set or announced, a TDP for Oracle could be available on Linux with TSM v5 in the first half of 2002. -- Tom Thomas A. La Porte DreamWorks SKG [EMAIL PROTECTED] On Wed, 24 Oct 2001, Joshua S. Bassi wrote: There aren't any TDP agents for Linux at this time. The best way to backup the Oracle data on Linux would be to backup or archive the Oracle dump. -- Joshua S. Bassi Independent IT Consultant IBM Certified - AIX/HACMP, SAN, Shark Tivoli Certified Consultant- ADSM/TSM Cell (408)(831) 332-4006 [EMAIL PROTECTED] -Original Message- From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED]] On Behalf Of ±èÀο± Sent: Tuesday, October 23, 2001 9:09 PM To: [EMAIL PROTECTED] Subject: oracle backup(linux) Is there tdp for oracle(linux)?? If so, what is the procedure for backing up oracle?? server: solaris 2.6 client: redhat 6.2 Thanks in advance.. CD DATA CORP. IDC TEAM. KIM IN YOP 14th Floor,Sam Jung Bldg. 237-11,Nonhyun-Dong,Gangnam-Gu, Seoul,Korea 135-831 TEL: 82-2-546-3108 MOBILE: 016 523-2032 FAX: 82-2-514-9007
Re: Expiration challenge
I can think of two options. 1) Save an additional backup of your (ADSM) database, and take measures to ensure that the tapes upon which your (Informix) database reside do not get reused. This will allow you to restore the ADSM database at some stage in the future, and the tapes that you care about with the Informix database will still be available from which you can restore the Informix OLTP data. 2) Restore a copy of the Informix OLTP database to some location, and then run an *archive* of that data. Based on your description of the environment, unless you can get a vendor to loan you equipment for the process, I would think that of the two options I've proposed, #1 is the better choice. I'm sure others on the list might approach this question from a different angle and offer further suggestions. -- Tom Thomas A. La Porte DreamWorks SKG [EMAIL PROTECTED] On Fri, 21 Sep 2001, Prose, Fred wrote: We run Tivoli 4.2 on AIX (4.3.3) with a primary emphasis on backing up Informix databases using the Informix TDP. We defined a database management class for the database backups, with retain extra versions set to 30 days. All database servers (13) go to this same class. This has always proved quite adequate, as the requirement has been to provide recovery capabilities, not archive historical copies of the data. We now have a Law Enforcement organization interested in some unusual activities that occurred in one of sixty databases that resides on one of the servers. The request they placed was to retain the oldest copy that we had of this database until they were through with their investigation. Restoring the database to a different server is not as simple an option as it first might seem. With an Informix database backup, you are backing up the entire OLTP environment, not just a single database. This entire environment would need to be restored to a different server and as luck would have it, the server in question happens to be the biggest and no other server could handle the restore (space wise). Saving the current environment, restoring from the backup, extracting the database in question, and then restoring the current environment I've estimated to take twelve hours. While it might be the only option, it's not a popular one. To prevent the expiration I set retention to 365 days on the management class. That will hopefully provide the opportunity to sort things out. The problem now is that without expiring versions, I'm not recycling tapes (as expected) and buying additional tapes during a budget crisis is a difficult and long process. Here's the questions: Since I know what tapes (2) the current backup is located on, is there a way to freeze those tapes so that the files contained on the tapes do not expire? If I set up a new management class for this one server, can I force existing files from this server into this class?
Re: Backup of EMC
Jeff, NDMP support is meant to be available at the end of this month. That they haven't set pricing information for this yet makes me wonder if it will actually be available any time soon. I'm skeptical, but hopeful, as we've several TB of NetApp storage in our environment. -- Tom Thomas A. La Porte DreamWorks SKG [EMAIL PROTECTED] On Thu, 6 Sep 2001, Jeff Bach wrote: IBM people, How does IBM/Tivoli plan to provide for this need in order to prevent customer from being required to implement hardware vendor solutions such as EDM to backup this data? The competition is direct fiber attaching to the data and backing it up. Jeff Bach Home Office Open Systems Engineering Wal-Mart Stores, Inc. WAL-MART CONFIDENTIAL -Original Message- From: Michael Bartl [SMTP:[EMAIL PROTECTED]] Sent: Thursday, September 06, 2001 9:01 AM To: [EMAIL PROTECTED] Subject:Re: Backup of EMC Mahesh, on NAS boxes like CLARiiON or NETAPP you won't find a backup/archive client that runs directly on the machine. To get your backup done just use another machine in your LAN that has enough network bandwith available to both the NAS box and the TSM server. On WinNT you can use the UNC-Name in dsm.opt: DOMAIN \\NASBOX\SHARE With Unix you have to define a mountpoint for the tree you want to backup. Then you put the mountpoint into your optionfile. Good luck, Michael Bartl Office of Technology, IT Germany/Austria Cable Wireless Deutschland GmbH. Landsberger Str. 155Tel.: +49 89 92699-806 80687 Muenchen Fax.: +49 89 92699-302 mailto:[EMAIL PROTECTED] http://www.de.cw.com Mahesh Tailor wrote: Hello, everyone. A group in our department just received an EMC CLARiiON system. On this system is a filesystem that I need to backup. How can this be done? I have never dealt with this beast. 3466 Network Storage Manager running TSM v3.7.4 and AIX 4.3.2. Thanks for any help and advice in advance. Mahesh Tailor WAN/NetView/TSM Administrator Carilion Health System Voice: 540-224-3929 Fax: 540-224-3954 ** This email and any files transmitted with it are confidential and intended solely for the individual or entity to whom they are addressed. If you have received this email in error destroy it immediately. **
Re: Transfer Rate
Also, if you're client and server are on the same host, the network shouldn't enter into the picture, you should be using SHAREDMEM as your protocol. -- Tom On Thu, 2 Aug 2001, David Longo wrote: Your original post of actlog stats didn't include how many objects inspected/backed up. If you have a LOT of small files that would account for some of it. What type of network do you have? I just looked at some of my stats and I have on the clients I checked a Network to Aggregate data transfer time of 2 -1 to 3 -1. How many sessions does the client use? You can set RESOURCEUTILIZATION in dsm.sys on client to increase number of sessions - see docs for info. Experiment to see best for you. I would think you can get better than 8 hours for 70GB. I imagine 4 -6 hours is as good as you can do. Have TSM 3.7.4.0 server on AIX 4.3.3 on F50 with 4x 332Mhz CPU and 2 GB RAM. IBM 3575 library. Using 155MB ATM network. (Some clients have 100MB Ethernet.) David B. Longo System Administrator Health First, Inc. 3300 Fiske Blvd. Rockledge, FL 32955-4305 PH 321.434.5536 Pager 321.634.8230 Fax:321.434.5525 [EMAIL PROTECTED] [EMAIL PROTECTED] 08/02/01 01:47PM The only client we have is where the server is located. We are running TSM to backup one RS/6000 F50 server. And backing up 70 GB in 8 hours doesn't seem right. Is there anything that I can do to speed this up. We run a Magstar 3570 connected to the F50 server through SCSI. TSM Version 4.1.2.0 AIX Version 4.3.3.0 -Original Message- From: Lindsay Morris [mailto:[EMAIL PROTECTED]] Sent: Thursday, August 02, 2001 10:19 AM To: [EMAIL PROTECTED] Subject: Re: Transfer Rate Aggregate uses wall-clock time; Network uses the intervals between when the client says I'm sending data now, to when the server actually receives it. It's perfectly normal to have aggregate be a lot slower than network. If you really want to speed up your backups, you need to look at the accounting log records (.../server/bin/dsmaccnt.log - see admin guide for layout) to see which of your clients suffer most from idle wait, media wait, or comm wait. -Original Message- From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED]]On Behalf Of Bill Wheeler Sent: Thursday, August 02, 2001 9:55 AM To: [EMAIL PROTECTED] Subject: Transfer Rate Hello All, I hope a few of the *SMer guru can help me out on this one. I don't know if you have had a problem like this before. The problem is our Aggregate Transfer Rate is allot lower then the Network Transfer Rate. Is there something that I can look at to see if I can get the data moved quicker?Here is a copy of the Activity log with the rates: ANE4961I (Session: 4532, Node: F50_CLIENT) Total number of bytes transferred:69.33 GB07/31/01 06:51:30 ANE4963I (Session: 4532, Node: F50_CLIENT) Data transfer time:11,858.22 sec 07/31/01 06:51:30 ANE4966I (Session: 4532, Node: F50_CLIENT) Network data transfer rate:6,130.86 KB/sec07/31/01 06:51:30 ANE4967I (Session: 4532, Node: F50_CLIENT) Aggregate data transfer rate: 2,462.14 KB/sec 07/31/01 06:51:30 ANE4968I (Session: 4532, Node: F50_CLIENT) Objects compressed by:0% 07/31/01 06:51:30 ANE4964I (Session: 4532, Node: F50_CLIENT) Elapsed processing time:08:12:07 Any advice would be helpful. Thanks in advance, Bill Wheeler AIX Administrator La-Z-Boy Incorporated [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] MMS health-first.org made the following annotations on 08/02/01 14:20:19 -- This message is for the named person's use only. It may contain confidential, proprietary, or legally privileged information. No confidentiality or privilege is waived or lost by any mistransmission. If you receive this message in error, please immediately delete it and all copies of it from your system, destroy any hard copies of it, and notify the sender. You must not, directly or indirectly, use, disclose, distribute, print, or copy any part of this message if you are not the intended recipient. Health First reserves the right to monitor all e-mail communications through its networks. Any views or opinions expressed in this message are solely those of the individual sender, except (1) where the message states such views or opinions are on behalf of a particular entity; and (2) the sender is authorized by the entity to give such views or opinions. ==
Re: Backing up lots of small files
Is it possible to break up the filesystem(s) with VIRTUALMOUNTPOINT options? This will let you break down the problem into smaller chunks, and possibly run your 'dsmc' processes out of cron, rather than via the scheduler. On Tue, 31 Jul 2001, Gerald Wichmann wrote: Has anyone come up with any good strategies on backing up a file server with millions of small (1KB-10KB) files? Currently we have a system with 50GB of them that takes about 40 hours to backup. Many of the files change.. I m wondering if anyone has anything in place that works well for this sort of situation.. Something like this or perhaps something I haven t thought of: develop script that will tar/compress the files into a single file and back that single file up daily. Perhaps even search the file system by date and only backup files that have changed since the last backup (this seems like it wouldn t be any faster then simply backing them up though)
Netapps for ADSM Server?
Just out of curiosity, has anybody experimented with using a Network Appliance, or other similar network attached storage devices, for their ADSM database or logs? We are using Network Appliance Filers very successfully on our Oracle servers, so I was curious about the possibility of using them on our ADSM servers. Just wondering if anybody else has gone down this path before me, or have even thought about it. -- Tom Thomas A. La Porte DreamWorks SKG [EMAIL PROTECTED]
Re: Data on my tapes
Jeff, Compressing more than once generally doesn't gain anything in terms of space reduction, in fact there are certain instances in which additional compression passes result in *larger* files. Since you are compressing at the client side, and you are averaging about 35GB on a 30GB native tape, I'd say you're getting pretty good storage utilisation. -- Tom Thomas A. La Porte DreamWorks SKG [EMAIL PROTECTED] On Wed, 20 Jun 2001, Jeff Bach wrote: What is my problem? I am just going to 3590K. I use client compression on everything. I compress on the tape drives also. I tried to define the format as 3590b, but it doesn't allow me to access the drives. I currently have it set to drives. Has anyone seen this? This seems to be a small amount of data. H00924DB_TPOOL TAPE3590K35,694.3 100.0Full H00925DB_TPOOL TAPE3590K35,490.6 100.0Full H00926DB_TPOOL TAPE3590K35,765.3 100.0Full H00927TAPE3590 TAPE3590K36,007.6 100.0Full H00928TAPE3590 TAPE3590K40,960.0 29.4 Filling H00929DB_TPOOL TAPE3590K35,635.3 100.0Full H00930DB_TPOOL TAPE3590K35,554.6 100.0Full H00932DB_TPOOL TAPE3590K40,960.0 49.2 Filling H00933DB_TPOOL TAPE3590K35,882.4 100.0Full H00934DB_TPOOL TAPE3590K35,904.1 100.0Full H00935DB_TPOOL TAPE3590K35,629.7 100.0Full H00936DB_TPOOL TAPE3590K35,887.2 100.0Full H00937DB_TPOOL TAPE3590K35,786.8 100.0Full H00938DB_TPOOL TAPE3590K35,710.0 100.0Full Jeff Bach -Original Message- From: Richard Sims [SMTP:[EMAIL PROTECTED]] Sent: Wednesday, June 20, 2001 10:28 AM To: [EMAIL PROTECTED] Subject: Re: Data on my tapes Ok, how much data are you getting on a 3590 K tape. David - A rough visual average that I see on my full tapes is 58 GB, in backup and copy storage pools. (If using HSM, expect much less, because it doesn't Aggregate.) Richard Sims, BU ** This email and any files transmitted with it are confidential and intended solely for the individual or entity to whom they are addressed. If you have received this email in error destroy it immediately. **
Upgrading to 4.1
Platform: AIX 4.3.2 ADSM: 3.1.2.50 We have been testing 4.1 on a system that already has a 3.1 instance installed. We are now satisified with 4.1, and would like to migrate our 3.1.2.50 instance to 4.1. The manual indicates directions on how to do the migration at the time of installing the 4.1 software--a step that we've obviously already taken. I thought that what I would have to do is 1) do a full database backup of the 3.1.0.5 instance 2) copy my relevant files (dsmserv.dsk, etc.) to the new DSMSERV_DIR (/usr/tivoli/tsm/server/bin) 3) start the TSM 4.1 server with 'dsmserv upgradedb' 4) re-register all of my licenses Has anybody else gone through the process in this manner? Am I missing something? Any advice/tips/suggestions are welcome. -- Tom Thomas A. La Porte DreamWorks SKG [EMAIL PROTECTED]
Re: Network Appliance
John, Is there a reference to a product announcement of any sort? Searching Tivoli's web site for NDMP, NAS, or TSM 4.2 does not turn up anything other than the future directions white paper that has been there for about a year which frustratingly discusses NDMP support in the present tense. -- Tom Thomas A. La Porte DreamWorks SKG [EMAIL PROTECTED] On Wed, 9 May 2001, John Bremer wrote: Geoff, Tivoli announced that TSM version 4.2 should provide backup and restore of NAS filers - 3Q 2001. Tivoli Data Protection for NAS is a specialized client that interfaces with the Network Data Management Protocol (NDMP). Full volume image backup/restore will be supported. File level support is announced for TSM version 5.1 - 1Q 2002. John At 08:16 AM 5/2/01 -0700, Gill, Geoffrey L. wrote: Is anyone familiar with Network Appliance products and can TSM back them up directly or any other way? Thanks, Geoff Gill TSM Administrator NT Systems Support Engineer SAIC E-Mail: [EMAIL PROTECTED] Phone: (858) 826-4062 Pager: (888) 997-9614
Re: Reducing/compressing the database
Is anybody else a little bit distressed that Tivoli's TSM administrator needs to seek this advice from the list, rather than from the developers of the product? This is an issue that has been one that has generated much discussion on and off with regards to the benefits and costs of attempting to reorganize a TSM database, and one on which we have never had much of a substantive response from anyone within Tivoli, either in this forum or in the documentation of these commands in the manuals. I would have thought that the people at Tivoli would be the best source for information of this type. I'm a little bit concerned that our collective guessing and anecdotal evidence is all that any of us as administrators have to go on. -- Tom Thomas A. La Porte DreamWorks SKG [EMAIL PROTECTED] On Sun, 22 Apr 2001, Shawn Drew wrote: TSM v3.7.3 I have read alot on this list about reducing the database because my situation is pretty bad. We have a 103 gig database that was 97% used! I finally was permitted to fix the outrageous retention settings, and got it down to 50% utilization, but the Maximum Reduction value is still 0. Now, I want to get the database's assigned capacity to the ibm recommended max of 70 gigs. (This is from the ISSS last year in san diego. I was the guy that had an 80gig db in one of the seminars) From reading this list, it seems I have a few options, but could not determine the best route. Down time IS a factor for this. The performance on this machine, for restores, is dramatically slower than on other machines, and since it seems all else is almost equal, I am assuming its the db size. First of all, my reason for doing this is to get better performance on my restores. So, will defragging the database really improve my restore times? seems pointless otherwise. It seems my options are: - dumpdb/loaddb - I read some horror stories of this, and really hesitate on it. Also, it seems the loaddb takes very long, from other people's experience (2 days for a 40gb db? I think I read?) - unloaddb/loaddb - The only difference I can find with this and the previous one is that it defrags the database. And it seems that the loaddb portion is the same, and is subject to the same unreliability and time problems. (This is the one described in the manuals to solve my situation) - Richard recommended the backup db/restore db options over the dumpdb/loaddb because it is more reliable and faster. Does this also offer the defrag benefit? And how long would it take? - Migrate all the client nodes one by one to other machines with import/export. Kill and rebuild TSM and the db and move everything back. This seems like the one with least downtime. which may be the best option and least risky. But it will take a LONG time, and strain my other boxes. thanks shawn ___ Shawn Drew Tivoli IT - ADSM/TSM Systems Administrator [EMAIL PROTECTED]
Re: TDP for Oracle
What about Linux? Is there plans for TDP for Oracle on Linux? This would be extremely useful, especially since Linux is now a tier one platform for Oracle. On Thu, 19 Apr 2001, Thiha Than wrote: hi Geoff, We don't have it planned in the near future. regards, Thiha Date:Wed, 18 Apr 2001 11:22:45 -0700 From:"Gill, Geoffrey L." [EMAIL PROTECTED] Subject: TDP for Oracle Does anyone know if TDP for Oracle on True64 is in the works, and if it is a projected release date? Geoff Gill TSM Administrator NT Systems Support Engineer SAIC E-Mail: [EMAIL PROTECTED] Phone: (858) 826-4062 Pager: (888) 997-9614
Re: How to find out what tape's a client data are on ?
Just out of curiosity, since I've only casually been following this thread, was there a specific reason for the "group by" clause? I think what is desired is simplly an "order by" clause, e.g. select volume_name, node_name - from volumeusage - where node_name = 'MEDRS1' - and stgpool_name = 'COPYPOOL' - order by volume_name, node_name The group by function is necessary when there is a group function in the "select" statement, e.g. when you are summing a value: select node_name, sum(backup_copy_mb) - from auditocc - where node_name = 'MEDRS1' - group by node_name -- Tom Thomas A. La Porte DreamWorks SKG [EMAIL PROTECTED] On Thu, 1 Mar 2001, Shekhar Dhotre wrote: Thanks David ... tsm: TSMselect volume_name,node_name from volumeusage where node_name='MEDRS1' and stgpool_name='COPYPOOL' group by volume_name,node_name VOLUME_NAMENODE_NAME -- -- MED200 MEDRS1 MED202 MEDRS1 MED251 MEDRS1 MED260 MEDRS1 MED261 MEDRS1 MED501 MEDRS1 MED507 MEDRS1 MED511 MEDRS1 MED513 MEDRS1 tsm: TSMselect volume_name,node_name from volumeusage where node_name='MEDRS1' and stgpool_name='TAPEPOOL' group by volume_name,node_name VOLUME_NAMENODE_NAME -- -- MED201 MEDRS1 MED255 MEDRS1 MED506 MEDRS1 David Longo [EMAIL PROTECTED]@VM.MARIST.EDU on 03/01/2001 04:28:55 PM Please respond to "ADSM: Dist Stor Manager" [EMAIL PROTECTED] Sent by: "ADSM: Dist Stor Manager" [EMAIL PROTECTED] To: [EMAIL PROTECTED] cc: Subject: Re: How to find out what tape's a client data are on ? Your select statement is incorrect. It should be: select volume_name,node_name from volumeusage where node_name='MEDRS1' and stgpool_name='COPYTAPE' group by volume_name,node_name David B. Longo System Administrator Health First, Inc. 3300 Fiske Blvd. Rockledge, FL 32955-4305 PH 321.434.5536 Pager 321.634.8230 Fax:321.434.5525 [EMAIL PROTECTED] [EMAIL PROTECTED] 03/01/01 02:48PM The following query shows same tapes for node 'MEDRS1' in tapepool and copypool ? tsm: TSMselect volume_name,node_name from volumeusage where node_name ='MEDRS1'g roup by volume_name,node_name where stgpool_name='TAPEPOOL' VOLUME_NAMENODE_NAME -- -- MED200 MEDRS1 MED201 MEDRS1 MED251 MEDRS1 MED254 MEDRS1 MED255 MEDRS1 MED260 MEDRS1 MED261 MEDRS1 MED501 MEDRS1 MED506 MEDRS1 MED507 MEDRS1 MED511 MEDRS1 MED513 MEDRS1 tsm: TSMselect volume_name,node_name from volumeusage where node_name='MEDRS1'g roup by volume_name,node_name where stgpool_name='COPYPOOL' VOLUME_NAMENODE_NAME -- -- MED200 MEDRS1 MED201 MEDRS1 MED251 MEDRS1 MED254 MEDRS1 MED255 MEDRS1 MED260 MEDRS1 MED261 MEDRS1 MED501 MEDRS1 MED506 MEDRS1 MED507 MEDRS1 MED511 MEDRS1 MED513 MEDRS1 tsm: TSMQ DRM Volume NameState Last Update Automated Date/Time LibName - --- 64 Vault 02/22/01 12:08:00 MED200 Vault 02/16/01 15:38:21 MED251 Vault 02/16/01 15:38:22 MED252 Vault 02/23/01 13:38:02 MED254 Vault 02/26/01 10:43:13 MED260 Vault 02/19/01 10:36:43 MED261 Vault 02/19/01 10:36:43 MED271 Vault 02/26/01 10:43:13 MED311 Vault 02/19/01 10:36:43 MED375 Vault 02/27/01 10:48:18 MED414 Vault 02/20/01 11:42:36 MED501 Vault 03/01/01 12:47:40 MED503 Vault 02/21/01 10:47:50 MED511 Vault 02/27/01 10:48:18 MED513 Vault 03/01/01 12:47:40 MED514 Vault 03/01/01 12:47:40 MED515 Vault 03/01/01 12:49:00 MED512 Vault 03/01/01 12:49:00 more... (ENTER to continue, 'C' to cancel) MED413 Vault 02/27/01 10:49:43 MED510 Vault 02/26/01 10:44:30 MED374
TSM Pricing [was Re: Performance Large Files vs. Small Files]
We, too, were a bit rudely awakened by this new pricing structure when we purchased our upgrade to 4.1. From my perspective, the most agonizing aspect of the structure is the need to pay for a tape library access license (I forgot what the "feature" is actually called) for every physical server that accesses a tape library, rather than for each tape library. We did the math, and it almost made more sense for us to purchase a new S80 class machine that could effectively handle multiple TSM instances, rather than purchase additional tape library access licenses to continue to run on our two separate F50 machines. That begins to make very little sense when the price point of a "feature" outweighs the cost of the hardware to implement that feature. Particularly a feature that is so fundamental to the running of the application: I can't imagine trying to run a large-scale ADSM implementation without a 3494 (or similar) tape library. -- Tom Thomas A. La Porte DreamWorks SKG [EMAIL PROTECTED] On Wed, 21 Feb 2001, bbullock wrote: That is also an option that we have not considered. It actually sounds like a good one. I'll bring it up to management. The only problem I see in this is the $$$ needed to purchase a TSM server license for every manufacturing host we try to back up. I don't know about your shops, but with the newest "Point based" pricing structure that Tivoli implemented back in November (I believe it was about then). They are now wanting to charge you more $$$ to run the TSM server on a multi-cpu unix host than on a single host NT box. In our shop where we run TSM on beefy S80s, that means a price change that is exponentially larger than what we have paid in the past for the same functionality. Ben Bullock UNIX Systems Manager -Original Message- From: Suad Musovich [mailto:[EMAIL PROTECTED]] Sent: Wednesday, February 21, 2001 4:37 AM To: [EMAIL PROTECTED] Subject: Re: Performance Large Files vs. Small Files On Tue, Feb 20, 2001 at 03:21:34PM -0700, bbullock wrote: ... How many files? Well, I have one Solaris-based host that generates 500,000 new files a day in a deeply nested directory structure (about 10 levels deep with only about 5 files per directory). Before I am asked, "no, they are not able to change the directory of file structure on the host. It runs proprietary applications that can't be altered". They are currently keeping these files on the host for about 30 days and then deleting them. I have no problem moving the files to TSM on a nightly basis, we have a nice big network pipe and the files are small. The problem is with the TSM database growth, and the number of files per filesystem (stored in TSM). Unfortunately, the directories are not shown when you do a 'q occ' on a node, so there is actually a "hidden" number of database entries that are taking up space in my TSM database that are not readily apparent when looking at the output of "q node". Why not put a TSM server on the Solaris box and back it up to one of the other servers as a virtual volume. It would redistribute the database to the Solaris host and the data is kept as a large object on the tape-attached TSM server. I also remember reading about grouping files together as a single object. I can't remember if it did selective groups of files or just whole filesystems. Cheers, Suad --
Re: Performance Large Files vs. Small Files
Imagine it strictly from a database perspective.
Scenario 1: 15 files, 2GB each
Scenario 2: 15728640 files, 2KB each
In scenario one, your loop is essentially like this:
numfiles = 15;
for (i = 0; $i $numfiles ; $i++) {
insert file characteristics into database;
request data be sent from client;
store data in storage pool;
}
In scenario two, the primary difference is that numfiles =
15728640:
numfiles = 15728640;
for (i = 0; $i $numfiles ; $i++) {
insert file characteristics into database;
request data be sent from client;
store data in storage pool;
}
This means that, in the first scenario, there are 15 interactions
with the database, 15 system calls on the client for file
open/read operations, etc. In the second scenario, there are 15
*million* interactions with the database, 15 *million* file I/O
operations, etc.
Realistically, this is a bit of a simplification, as the
TXNGROUPMAX and the TXNBYTELIMIT parameters help to group the
files into transaction batches that can be larger than a single
file, which reduces the number of round trips to the database,
but the overall effect is still there.
Although you may be transferring the same amount of aggregate
data, you have to factor in the overhead of each single
transfer. Although the overhead may be small, if you multiply
that small number by two orders of magnitude you do generally end
up with a big number.
Imagine the time it would take to collect $30 million dollars
from fifteen $2 million donors, then think of collecting the same
amount of money from fifteen million $2 donors.
I would recommend that you break up your NT server into smaller
filespaces, either physically on the NT server, or logically with
virtualfilespaces on the ADSM server. That way you can have more
multiple processes working simultaneously on the backup. The
aggregate time it will take to back up the server will be the
same, but the wall clock time will be approximately divided by
the number of processes you can run simultaneously.
-- Tom
Thomas A. La Porte
DreamWorks SKG
[EMAIL PROTECTED]
On Wed, 14 Feb 2001, Diana J.Cline wrote:
Using an NT Client and an AIX Server
Does anyone have a TECHNICAL reason why I can backup 30GB of 2GB files that are
stored in one directory so much faster than 30GB of 2kb files that are stored
in a bunch of directories?
I know that this is the case, I just would like to find out why. If the amount
of data is the same and the Network Data Transfer Rate is the same between the
two backups, why does it take the TSM server so much longer to process the
files being sent by the larger amount of files in multiple directories?
I sure would like to have the answer to this. We are trying to complete an
incremental backup an NT Server with about 3 million small objects (according
to TSM) in many, many folders and it can't even get done in 12 hours. The
actual amount of data transferred is only about 7GB per night. We have other
backups that can complete 50GB in 5 hours but they are in one directory and the
# of files is smaller.
Thanks
Network data transfer rate
--
The average rate at which the network transfers data between
the TSM client and the TSM server, calculated by dividing the
total number of bytes transferred by the time to transfer the
data over the network. The time it takes for TSM to process
objects is not included in the network transfer rate. Therefore,
the network transfer rate is higher than the aggregate transfer
rate.
.
Aggregate data transfer rate
The average rate at which TSM and the network transfer data
between the TSM client and the TSM server, calculated by
dividing the total number of bytes transferred by the time
that elapses from the beginning to the end of the process.
Both TSM processing and network time are included in the
aggregate transfer rate. Therefore, the aggregate transfer
rate is lower than the network transfer rate.
NDMP support?
Is there any sense that TSM might support NDMP anytime soon? The single reference to NDMP is at http://www.tivoli.com/products/documents/whitepapers/storage_vision.html, which is over a year old, yet contains frustrating present-tense descriptions of production functionality that still doesn't exist. Just wondering if Tivoli is going to either update their vision, or include any of the myriad functions discussed in this vision paper in an actual product. -- Tom Thomas A. La Porte DreamWorks SKG [EMAIL PROTECTED]
Re: Generting Commands from a Script
select 'move data ' || volume_name - from volumes - where upper(stgpool_name)='LNTAPECO' - or upper(stgpool_name)='LNTAPENO' - or upper(stgpool_name)='UNTAPECO' - or upper(stgpool_name)='UNTAPENO' - or upper(stgpool_name)='NTTAPECO' - or upper(stgpool_name)='NTTAPENO' and pct_utilized 50 You'll probably need to set sqldisplaymode to wide, or if you run it in batch to commadelimited or tabdelimited. Redirect the output to a file and you've got a macro file that you can run. -- Tom Thomas A. La Porte DreamWorks SKG [EMAIL PROTECTED] On Wed, 24 Jan 2001, Joseph Marchesani wrote: Group I have created a script name VOLUME_PCT with the following results: ANR1461I RUN: Executing command script VOLUME_PCT. ANR1466I RUN: Command script VOLUME_PCT, Line 5 : select volume_name, stgpool_name,pct_utilized from volumes where upper(stgpool_name)='LNTAPECO' and pct_utilized 50 or upper(stgpool_name)='LNTAPENO' and pct_utilized 50 or upper(stgpool_name)='UNTAPECO' and pct_utilized 50 or upper(stgpool_name)='UNTAPENO' and pct_utilized 50 or upper(stgpool_name)='NTTAPECO' and pct_utilized 50 or upper(stgpool_name)='NTTAPENO' and pct_utilized 50. VOLUME_NAMESTGPOOL_NAME PCT_UTILIZED -- -- A40004 NTTAPECO8.1 A40012 UNTAPENO2.0 A40015 NTTAPENO 48.0 A40021 NTTAPENO 32.9 A40081 NTTAPECO 47.4 A40156 NTTAPECO 15.4 ANR1494I RUN: Command return code is 0. ANR1487I RUN: Command return code is RC_OK (OK). ANR1462I RUN: Command script VOLUME_PCT completed successfully My question is how can I generated MOVE DATA commands using the VOLUME_NAMEs in the results of the above SCRIPT Thanks Joe Marchesani .
Re: periodic ADSM/TSM shutdown
We have never seen any issues with long running ADSM servers (note we're further down level than you are!): tsm: DLADSMq stat ADSM Server for AIX-RS/6000 - Version 3, Release 1, Level 1.5 Server Name: DLADSM Server Installation Date/Time: 08/06/1996 17:18:36 Server Restart Date/Time: 05/21/2000 20:58:39 [...] Folks who recommend periodic reboots and restarts are offering solutions that simply mask or alleviate any problems that may exist. On Thu, 12 Oct 2000, Pat Wilson wrote: I've never seen any evidence that a periodic restart is needed. Pat Wilson Dartmouth College [EMAIL PROTECTED] I am running AIX 4.3.3 on an S80 ADSM 3.1.2.40 (yes I know I am trying to get us to upgrade but you ever see anything move fast in corporate world?) A couple of weeks ago I ran into a problem that the server couldn't successfully backup the database and my log file filled up...I solved it by cycling the ADSM server. Then two days later BMC told us that we need to cycle the ADSM server on a regular basis. What are everyone's thoughts on that? I looked around and haven't seen any recommendations. TIA Becky Davidson Data Manager/AIX Administrator EDS/Earthgrains voice: 314-259-7589 fax: 314-877-8589 email: [EMAIL PROTECTED]
Re: H80 - TSM - 3494 --- Bottleneck?
Ronnie, Not sure I'm reading your message correctly, but my understanding was that the 3494 can connect to a host either via Ethernet or Serial cable, not via SCSI. The 3590 drives are attached to the host via either SCSI, or in your case fiber. Nevertheless, if your question is will the ethernet connection from the host to the 3494 be a bottleneck, the answer is no. The only traffic that passes across that connection is robotic instructions (e.g. mount tape A in drive 1, dismount tape B from drive 2, etc.). The actual data transfer between host and tape drive goes over the Fiber connection between the host and drive. -- Tom Thomas A. La Porte DreamWorks SKG [EMAIL PROTECTED] On Thu, 10 Aug 2000, Ronnie Crowder wrote: We are in the process of getting a new TSM server and library. I was thinking about how the hardware connects up to the 3494 and drives. I the drives are fiber connected to the rs6000 and the 3494 is scsi connected. This is what my question is: Do all the clients go through the single ethernet on the 6000 to get to the 3494 drives?? If so, is this a potential bottleneck and has anyone had any issues with this setup?? About how many clients are connected??
