Search results

In performance tests, I found that BiBlocksize of 16kb offered the best performance. If you increase your blocksize, you might want to lower your clustersize. Curious - how long does the db log show that it takes to backup the BI data ? What was going on at 14:11:01 ? The disk sync time looks...

Service Pack 3 in 10.1B offered some significant improvements. You should check the $DLC/version file to make sure you are at least on 10.1B03.

My problem was that I didn't specify the indexarea on the initial auditreconfig command. Because tables and indexes were in the same SA and they were both going to the same new SA, I didn't specify the indexarea in the command. After the first command was run incorrectly, I could not get the...

Linux - Red Hat 6.7 OE 11.5.1 - 64 bit Has anyone used the auditreconfig utility ? We use auditing on several large databases. We occasionally have a data fix or bug in a program that would generate 40 GB or more of audit data. After the audit data is archived, we are left with large amounts...

Progress created a couple items to help with OS backups - proquiet and ai sequence. We use OS backups with proquiet to backup large databases that would take probkup several hours to complete. You can / should restore databases taken with OS backups so you can ensure your backup strategy is...

We run a couple databases with clustersize of 256 and a few more at 128. The only downside I am aware of is the startup time when the db formats the 4 bi clusters. I would be curious to know the total time of the checkpoint and then how much of that was the sync time. promon R&D 3 4 (I think)...

Do you compress the ai log before transferring to DR ? You can check the db log of the DR database to see where the time is being spent - is it processing the BI or AI. Keeping your active transactions as small as possible helps apply AI logs faster. We have some character screens in our...

We have some automation around sending USR1 signals on Linux. It works better for us than using higher levels of logging. We use lsof to find the pid's current working directory and then send the USR1 signal, grab the protrace file and send the important details in an email or alert. Works...

Seems like binary D&L couldn't load LOB into a different area number than it was dumped from in 10.x That might have been fixed in later versions. I believe we renamed the original table, created the new table with LOB data going to a separate area and then used raw transfer to move the data over.

Maybe I haven't added an index to a large enough table, but for our TB+ database it is a lot faster to add the index inactive and then activate while AI remains active than to go through the process of taking the extra backup and shipping it to DR. I see you can take a backup of a TB database in...

Do you know if the database is Enterprise or Workgroup also what version number you are using ?

As suggested, the _UserTableStat and _UserIndexStat would be a good place to start looking. You will probably have to set up the DB so more than the first 50 tables will be monitored and that will require a restart. If these can identify a table that is read millions of times then it would be...

its SIGUSR1 that generates the ABL stack - I don't know what USR2 would do. The active clients (now, peak) isn't very helpful for us. I concentrate on how many agents are available. The lines in the logs you have highlighted seem to be normal lines - I don't enable trace logging on the app...

Your buffers flushed seems way to high if you are hitting a checkpoint every couple minutes. Were you able to check the sync times on the checkpoints ? adding the memory buffers should help reduce the burden on the disk.

I like to look at the Sync times for the database checkpoints - R&D 3 4 I like to look at the bi stats - R&D 2 5 I like to know how many disk reads / sec the db is doing. 5 - summary screen. sometimes people use the bigrow test on an empty database to see how long it takes the disk to allocate...

The UUID fields were added to existing tables. Yes the UUID field caused the record to fragment. I don't recall the fragmentation factor - I know that in a table with 100 million records, I had 15 million or so additional fragments. Yes a D&L resolved the issue. Thanks Rob - so it was the...

reads from disk increased as did logical reads. The reads from disk is what was impacting our performance as the disk io was not stellar. table analysis showed the number of fragments was increased while the number of records remained the same. the UUID fields were implemented as a way to...

assuming you didn't copy the scripts to /usr/bin as part of the install. Stop all processes using 11.3 Delete the install directory Delete the work directory Run the install again. If scripts were copied to /usr/bin, then those will need to be manually removed if they aren't wanted there.

We had problems with fragmentation on Type II storage areas, when we added UUID fields to many highly read tables that didn't have a unique identifier. Our reads noticeably increased. We don't go above 128 RPB and we were using 1 RPB for index areas, but I think 8RPB was recommended for...

Roll forward performance isn't much better in 10.2B, it will be a few months before I know what 11.5 roll forward performance is. One thing you can check is to make sure your DR server has the same disk performance as prod. Often companies deploy cheaper - slower - disk in DR.