Indexer Not (indexing, that is...)
Last Friday I downloaded the new JBL archive resources. I noticed right away that they did not seem to be indexing. Searches performed would find appropriate material in all other resources, but not In any JBL resource. After waiting for a day with no indexing — I stopped and started the program several times to try to get it to index the new material, to no avail — I decided to be proactive and so issued a 'rebuild index' command. This seemed to work (i.e. I watched as the 'xx%" amount increased, as reported by the indexer, when I checked on it from time-to-time). However, after it was complete I found that I apparently have no indexes at all. A search for 'Moses' or 'Jesus' returns 0 hits in my library of over 1800 resources. So last night I did another 'rebuild index', which again appeared to proceed normally, or at least was preceding normally when I went to bed. This morning I find the same situation: no search returns any hits from my library. So, as a result, I have now started my third manual 'rebuild index'. As I write this, the indexer reports that it is 3% complete, and I look forward to a joyful day of work with the indexer churning away in the background.
I am not attaching logs at this point since I want to run this rebuild with me watching. Assuming that the results are the same as before, I will submit the logs later. In the meantime, any ideas about what may be going on, or not going on, as the case may be?
– DLA
Comments
-
David Adams said:
I am not attaching logs at this point since I want to run this rebuild with me watching.
Good idea; also could watch Activity Monitor.
David Adams said:In the meantime, any ideas about what may be going on, or not going on, as the case may be?
Couple possibilities: indexing crashed earlier or computer shutdown cancelled indexing. Without logs, not know what is going on.
Keep Smiling [:)]
0 -
OK ...
The indexing process seemed to proceed normally. I watched it for several hours until I had to teach a class. When the class was done I looked and saw that the indexer had quit, and I still did not have functioning indexes. I am certainly no expert, but I took a look at the logs (attached below), and it appears that the indexer went into 'download only' mode and terminated the indexing process. Here are the lines from the indexer.log:
05/09/2011 13:59:26 | INFO | 8 | IndexerProgram | Received check command event; processing command.
05/09/2011 13:59:26 | INFO | 8 | IndexerProgram | Next command is EnterDownloadOnlyMode.
05/09/2011 13:59:26 | INFO | 8 | IndexerProgram | Entering download-only mode.
05/09/2011 13:59:26 | INFO | 8 | IndexerProgram | Discovery or indexing was running; canceling it.
05/09/2011 13:59:26 | INFO | 8 | MacIndexerProgress | Stopping IndexerProgress timer.
05/09/2011 13:59:26 | WARN | 7 | LogosResourceIndexer | Ignored 383 invalid references for 'sesb-morph+he' in LLS:1.0.204 (out of 57813 total; 0%).
05/09/2011 13:59:26 | INFO | 7 | MacIndexerProgress | 1,081 of 1,834 resources indexed; current phase progress is 44.6%
If there is anyone at Logos who can help with this, I would certainly appreciate it. I will, as advised, 'keep smiling', but perhaps with clenched teeth.
– DLA
1385.LogosLogs David L. Adams 20110510-062608.gz
0 -
The Indexer stops when there is a download and if there is a failed resource download, which may be the case here due to time-out problems in the current beta, the indexer is not able to complete. We are hoping to have the download issue fixed soon, but I don't know if a fix will make it into the next beta or not. You could try working offline or disable automatic downloads temporarily to see if you can get the indexer to complete.0
-
I will disable automatic downloads and try it. Will let you know later what happens.
– DLA
0 -
With incremental indexing always being merged, wonder about modifying code that stops indexing when downloading ? (albeit downloading while indexing does increase disk contention a bit, especially while merging that may impact request time outs)
If indexing happening and downloads are complete, wonder about waiting for indexing to complete before showing dialog box to install new resources (since clicking OK will cause indexing to start over).
Looking at logs, noticed many network sync errors, followed by connection that found updated resources to download. At that time: 1,082 of 1,834 resources had been indexed.
Observation: previously have seen indexing cancelled by downloading (typically only few resources indexed).
Keep Smiling [:)]
0 -
Turning off automatic downloads worked, and the indexer ran to completion, so I have my indexes back. Happy, happy ... joy, joy!
But seriously, I think that KS4J is right to suggest that modifying the indexer code is in order, though I would recommend a different solution. Given the relative value of indexes to the program and the frequency with which download checks are done, it would make sense to me to defer any automatic calls to download while the indexer is indexing. Let the indexer finish first, then do an auto check for updates. This relatively minor modification would prevent this kind of problem from recurring in the future.
Thanks for the help.
–– DLA
0 -
Considering number of recent download issues (e.g. time outs), wonder about indexer discovery of new resources (from previous download session or scan command), indexing (with automatic merge), then download updates seems something worthwhile to consider. Could become an acronym: DID (Discover, Index, Download).
Thought about new installations: initially nothing to discover and index, so download (or scan) needed.
Keep Smiling [:)]
0
