On Tue, 08 Oct 2002 at 19:00:52 -0500, Searcher wrote:

> >>Also, after fter 4 days of running, it seems to have stopped now
> >>but it does not seem to be completed, just dead. What is the best way of 
> >Are your sure, what does aspseek/logs.txt and indexer output tell
> >you?
> 
> I've now noticed that it does not seem to be totally dead, because I see the 
> drive usage going up, extremely slowly mind you. Here are the last 20 lines of 
> the log. What should I be looking for here?
> 
> Got next   1002 URLs for:   0.490 seconds. Queued docs: 2525090.Time 
> 1034051731-1034051824.

Two things; first, lines that read 'Got next xxx URLs for: x.xxx seconds ...'
tell you that the queuer is still queueing documents in reasonable time frames
i.e. the first line means that 1002 URLs were added to the indexing queue (this
is indexers in memory runtime queue not the SQL db) in .490 seconds.  This also
tells you how many total URLs are sitting in the indexer runtime queue, at that
time it was 2,525,090 URLs.


>      Sec Count    Ch   Ch1   Ch2   New       Size       HQ  Hr hits HR lost  W 
> hit W miss  W ins
>  100.537   327   293   293     0   293    4416577    13863    12321     761  
> 49086   5621    396

Second bit of information comes from 'hit W miss  W ins ...' lines; these stats,
which are dumped every ~ 100 seconds, tell you various things about what the
indexer did over the previous ~ 100 second period and how well the internal
indexer caches are performing.  The fields have the following meaning:

Sec
        Period over which stats were gathered (seconds)
Count
        Total number of docs, received
Ch
        Total number of docs, where changes found from
        last indexing
Ch1
        Total number of docs, for which server did not
        return status 304
Ch2
        Total number of docs, which had non-empty
        "Last-Modified" for which server did not return
        status 304
New
        Total number of docs, which changed status from
        0 to non-zero
Size
        Total sum of document sizes, received
HQ
        Number of HREF cache queries
Hr hits
        Number of HREF cache hits
HR lost
        Number of URLs removed from HREF cache
W hit
        Number of WORD cache hits
W miss
        Number of WORD cache misses
W ins
        Number of WORD cache inserts

So, it does look like your indexing session is still working just fine.  To safely
terminate it you should use 'index -E' and then run 'index -D' to build the deltas.


Matt.

Reply via email to