Open Mic Q&A: All about Notes databases - 16 March 2011
Tech Exchange; STEwebcast; openmic call; open microphone call; discussion; conference call; customer call; ODS; NSF; functionality; issues; problems; submit; ask the experts; discuss; server; 8.5.1; open mic; questions; answers; questions and answers; q&a; replay; recording; webcast; FAQ; best practice; transcript; March 16th
IBM will hosted an Open Mic conference call with Lotus Development and Support Engineers on 16 March 2011.
Questions and Answers from the session:
Q: Is there a document which lists all ODS versions and features of each?
A: In the Notes/Domino Upgrade Cookbook -> Deploying Domino 8.5.x servers page -> Enabling new features section, we have a breakdown of the features/benefits of ODS 48 and 51 and how to enable each.
Q: Is there a way to search a server and find all templates by the "Template Name" in its database properties?
A: Look in the catalog.nsf
Q: Does the Automatic Corrupt Database Collection take corrupt databases off-line? What does the user get presented with if they try to access a corrupt database?
A: The Corrupt Database Collector does not take corrupt databases off-line. On database open, when the automatic capture of corrupt databases is enabled, a copy will be made of databases that are marked corrupt. The database will be left in place to be automatically or explicitly fixed up.
What a user sees will vary. If you have the database open and do not encounter any corruption, everything may work as you expect. If your activity takes you into the corruption, you will see failures. If you do not have the database open, then you may find you cannot open it because a FIXUP is being run on the database. You may also get an error opening the database if another user has the database open. If you are unable to run FIXUP on the database or if automatic FIXUP is enabled for the database, it may be taken offline, fixed up and brought back online.
Q: At Lotusphere 2011, there was some indication that changes might be coming regarding separating view indexes from the NSF, can anyone talk about that?
A: We are always looking to add value and reduce TCO for Domino. Removing view indexes from the main NSF containing the data documents and putting them into another NSF store in a potentially different location than the main NSF file is being researched as to the measurable benefits. There has been no commitment yet for this enhancement in RNext, but is still under research/discussion.
Q: When I run maintenance tasks on Notes databases, the transaction logs will be written and I receive an error message that the "Log File is Full". Will this also happen if I bring the server offline to run the maintenance tasks?
A: If you are seeing log file full errors, then you should open a PMR. Running the maintenance tasks with the server down may reduce the chance of a log file full, but depending on the root cause, there are no guarantees you won't see one.
Q: How can we upgrade the folder design (ie - load convert -u) while the mail files are 99% usage but we are not allowed by corporate policy to relax the existing mail quota? Are database quotas hard number, even preventing any server maintenance tasks?
A: The maintenance tasks will still run
Q: Is there a server command that will identify and record all of the template versions of mail files on a server?
A: You can build your own view in the Catalog.nsf to show all files by the templates. You can also look in the DbInheritTemplateName and DbTemplateName fields.
Q: My Notes mail file (based on 8.0.1 template) occasionally reverts to the 6.5.4 design. It only ever happens in one server replica at a time so deleting the icon from the workspace and 'finding replica' on its cluster server resolves the display problem. This has happened about three times over the past 3-4 months. What could be the cause?
A: There must be a 6.5.4 template on the server somewhere that is being used to replace the design.
Q: Are there any known issues when upgrading custom Domino applications from version 7.x to 8.x?
A: This is a question that we get when upgrading any version to any version. Make sure that you are running the latest release as there are always improvements to bugs and performance. You can check for specific bug fixes in the Fix List database. Every custom application is completely unique and it may work one way in the basic client and another way in the standard client or on the web. You must test all your custom apps because there is really no way for us to predict how your custom app might behave in a different version, given the infinite combination of design elements.
Q: We get an error message when setting the Out Of Office agent "Only plain text can be used in this type of field". We thought there was a problem with the OOO profile, so we replaced that, all forms, all subforms, and all database scripts, but we still get the same error on user's mail file. Any suggestions?
A: We have not seen that error before. Is your mail file customized, and if so, in what capacity? One guess is that someone is trying to paste in an emoticon or picture into a text field.
Q: Are there any tools/methods to remove '"local access protection" on NSFs that were provided from an outside source when we were not provided the .ID file and/or password?
A: There is not a way to access local access protection without the ID file. It must be decrypted before sending it to view it locally.
Q: Is there a way to access and detach a "File" resource from an NSF using LotusScript running on a Notes client?
A: Yes, you can use the LS NotesNoteCollection to get handle to any design element, including resources.
Q: What server task is responsible for purging deletion stubs?
A: This action happens on database open.
Follow-up Q: Does the purge use a non-public API call? I've found that getModifiedNoteTable can return an error if there is a really huge number (3+ million!) of stubs and I don't know of any other way to get the stubs.
A: The purge happens via an internal call and we check once a day to see if we should try a purge. The getModifiedNoteTable failure is due to the fragmentation in the note ids being returned. Currently we don't have an option to work around that.
Q: Is there any issue with compacting log.nsf of the server? It increases and we need to recreate it in order to recovery its percentage used. We use 8.5.2 FP1 and 8.5.1 FP2.
A: You can run in-place compact with the server running without issue. Or you can run compact with the server down using ncompact.
Q: We are migrating from 7.0.3 to 8.5.2. Regarding overall capacity document counts, are they handled better in 8.5.2? We currently seem to have a lot of corruption using version R7.
A: In general, as we go through each release, we improve our corruption detection. We never expect when you move from one major release to the next that we are going to corrupt data. We would encourage you to move from R7 to 8.5.2FP1.
When upgrading from 7.x, there are two important disk space saving features:
1. Data compression was introduced in 8.0.1. You have to be at minimal ODS 48 to enable data compression. There are two options. One is to compress just the design documents and the other is to compress the data documents. We use LZ1 compression on the body of documents. We see 30-50% disk savings on ODS 48.LZ1 should be enabled on attachments as well.
2. In 8.5 we introduced Domino Attachment and Object Service (DAOS). DAOS saves significant space at the file level by sharing data identified as identical between databases (applications). For more information, see The new Lotus Domino attachment and object service.
Q: We are on version 6.5 and moving our servers to 8.5.1. Are there any issues with running ODS 43 for an extended amount of time on an 8.5.1 server?
A: ODS 43 is supported and we are not aware of any issues. Administrators like the fact that they did not have to upgrade the ODS at the same time during a major upgrade. However, you would not get the performance features until you upgraded to the new ODS version.
Q: When an attachment is in DAOS and a database replicates to a cluster member, if there are multiple databases with the same attachment, is the attachment only sent once to the clustermate?
A: DAOS is independent whether or not it is participating in a cluster. In the case of when the destination has DAOS enabled, the replication will phone ahead that you already had this attachment in your DAOS repository. If so, the attachment will not be retransmitted.
Q: My mail file is not being compacted when using compact -c with the 8.5.1FP5 server running.
A: On occasion, when running compact -c with the server running, and new mail arrives for that mail file while the database is being modified, that can cancel the compact operation. Try to bring the server down and then compact the database to see if the size shrinks as we would expect.
Q: We have concerns about old replicas polluting our databases after the deletion stubs are purged. Are there any best practices for protecting our databases from such occurrences?
A: There is a new replication option feature coming in Domino 8.5.3 called Purge Integration Replication Control (PIRC). PIRC is a replication option that will prevent documents from replicating back into a database older than the purge interval.
Q: Is there a way to analyze a database to reduce the index size?
A: You can issue the command "show database <database name>" on the server console to see all of the views that are currently built and their size in the database.
Q: We recently upgraded 8.0.1 to 8.5.2 on the client. We run FIXUP, UPDALL, and COMPACT once a week. We have a couple of users that never see any free space when deleting mail from the trash and archiving documents.
A: This should be handled automatically. With the server running, try a compact -B command. This will have a better chance of succeeding. There are also hotfixes in later fix packs for issues that started in 8.5 using compact -B.
Q: We see that some NLO files are deleted but we are not sure why they are being deleted. All of the messages get sent to an archive database which is DAOS enabled.
A: As the documents are migrated to the archive NSF, the reference count should go down, as they are removed from the primary NSF, but should go back up as they are deposited in the archive. It should end up being the same number of references at the end.
Q: We have two clustered 32-bit 7.0.4 servers on Windows 2003 server. Our goal is to move to two clustered 64-bit servers on Windows 2008. What is the best practice for moving databases over to the new system considering that I have two new servers?
A: Because you are running Windows 2003, which is supported with Domino 8.5.2, I would upgrade the servers in place and make sure the upgrade is successful. When you are ready to move to the new hardware, bring one of the clustermates down and copy everything over into the same location on disk to your current drive mapping. Then rerun the installation once all of the data is moved.
Q: We are going to change our Notes topology infrastructure from decentralized to centralized. I want to put all of our Domino 8.5.1 servers in the data center. We are concerned about the backups for information. What recommendations do you have for running compact before I backup all of the information?
A: You could upgrade to ODS 48 and take advantage of data compression to get the files to a smaller size. The backups would be a lot easier using data compression. It is always a good idea to bring the machines down and run FIXUP and initialize one last backup and then move the data back into the data center. For running compact, using compact -d will discard the views and -n for design and -v for data compression.
Q: We have a clustered environment and need to run compact -c. Is there another way to take a database offline without bringing down the server?
A: For clustered servers, our recommended best practice is to bring the server down and then run compact. You will get through multiple compactions faster with the server down. However, in the Admin client there is an option to take the database offline.
Q: All of our mail users have Designer access but our users are not able to run compact by going into application properties. The compact button is greyed out. Why is this occurring?
A: This is by design, but we have gotten the request before for users with Designer or Editor access to be able to run compact. We have had discussions about users having quotas assigned and if someone is near quota and has freed up space, compact should happen automatically on the back-end for them without any intervention.
For upcoming events, or Q&As from past events, see the Lotus Technical Exchange Webcasts page.