Frustrated with Tivoli Directory Server’s error messaging!

After quite a break from IBM Tivoli Directory Server (and deep hands-on technical stuff in general), I helped one of my teammates set up multi-master (aka peer-to-peer) replication between two ITDS servers. Setting up replication was a breeze once we had the two LDAP servers up and running, but getting to that initial state turned into bit of a nightmare. We initially tried to take a shortcut to setting up replication by importing a LDIF into both the LDAP servers that had all the replication entries pre-configured. However, I think there was an issue with the LDIF, and the shortcut did not work out. We then edited the LDIF file to remove the replication related entries, and tried to import the LDIF into ITDS using the ITDS Configuration tool after recreating the LDAP DB2 database. We had done this a couple of times with the LDIF that contained the replication entries with no issues, but when we attempted to bulkload the LDIF without the replication entries, the import failed, and all we got was an error message to the effect that the file was in use by another process. That error message totally threw us off! We were unsure about what file was in use thereby resulting in the error message and causing the bulkload to fail. Our suspicions centered around the LDIF file and we closed all processes that could have the LDIF file open, and based on a suggestion from my coworker, we tried searching for an open handle to the LDIF file using the “Sysinternals Process Explorer” tool. However all of this was in vain, and then our suspicion fell on the LDIF file being invalid syntactically. Tried a different LDIF file but got the same error message, and we figured that the file was ok. After wasting more time on this, we tried a brand new and simple ldif file with only a couple of entries. The bulkload for that file actually worked but we still got the same error message about “file in use by another process”. So, in the end, the “file in use” error message turned out to be a red herring – we were getting that message on both a success and a failure during bulkload – so much for error messaging! 😦

Not so good, IBM!