lots of log: "info DFM::findAll(): extent ... was empty, skipping ahead ..."

[老王]

my mongodb is newest stable version. log show lots of msg like bottom:

Wed Jul 6 16:05:06 [conn67475] info DFM::findAll(): extent 0:2900 was
empty, skipping ahead user_data_analysis.job
Wed Jul 6 16:05:07 [conn67475] info DFM::findAll(): extent 0:2900 was
empty, skipping ahead user_data_analysis.job
Wed Jul 6 16:05:07 [conn67475] info DFM::findAll(): extent 0:2900 was
empty, skipping ahead user_data_analysis.job
Wed Jul 6 16:05:07 [conn67475] info DFM::findAll(): extent 0:2900 was
empty, skipping ahead user_data_analysis.job

I try to convert index to a sparse index, but useless. I try to
reIndex index, same result.

please tell me WHY.

[Scott Hernandez]

So you are using 1.8.2? Can you run db.version() to verify that.

This messages is not an error. It is just informational.

Have you deleted lots of stuff? You may want to run a repair to
compact the data. This message is basically saying that there are
extents which are empty which are being skipped. This means the data
files are sparse; over time as you add more data they will get filled
back in and these messages will go away if you don't compact.

> --
> You received this message because you are subscribed to the Google Groups "mongodb-user" group.
> To post to this group, send email to mongod...@googlegroups.com.
> To unsubscribe from this group, send email to mongodb-user...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/mongodb-user?hl=en.
>
>

[老王]

Yes, I'm using 1.8.2. when system show a lots of MSG: "... was empty,
skipping ahead" , mongostat show result bottom:

insert query update delete getmore command flushes mapped vsize
res faults locked % idx miss % qr|qw ar|aw netIn netOut
conn set repl time
14 71 0 34 39 21 0 21.9g 50.4g
6.82g 0 33.1 0 0|0 1|1 18k 20k 624
IMOBILE2 M 10:51:40
8 84 0 39 41 21 0 21.9g 50.4g
6.82g 0 39 0 1|1 2|1 19k 24k 624
IMOBILE2 M 10:51:41
12 104 0 50 56 20 0 21.9g 50.4g
6.82g 0 49.1 0 0|0 2|0 24k 25k 625
IMOBILE2 M 10:51:42
8 187 100 43 79 9 0 21.9g 50.4g
6.82g 0 43.7 0 0|0 1|0 43k 228k 622
IMOBILE2 M 10:51:43
9 91 1 42 44 22 0 21.9g 50.4g
6.82g 0 40.8 0 0|0 2|0 20k 41k 622
IMOBILE2 M 10:51:44
8 95 0 47 49 11 0 21.9g 50.4g
6.82g 0 45.5 0 0|0 1|0 21k 22k 622
IMOBILE2 M 10:51:45
30 74 2 36 45 21 0 21.9g 50.4g
6.82g 0 35 0 0|0 2|0 25k 28k 618
IMOBILE2 M 10:51:46
24 72 0 36 45 26 0 21.9g 50.4g
6.82g 0 36 0 0|0 1|1 21k 31k 616
IMOBILE2 M 10:51:47
16 80 0 39 46 20 0 21.9g 50.3g
6.82g 0 37.8 0 0|0 2|0 20k 28k 615
IMOBILE2 M 10:51:48
15 104 1 51 55 21 0 21.9g 50.3g
6.82g 0 48.3 0 0|0 1|1 25k 29k 615
IMOBILE2 M 10:51:49

the write ops num aren't high, bug the locked% num is very high

On Jul 6, 8:46 pm, Scott Hernandez <scotthernan...@gmail.com> wrote:
> So you are using 1.8.2? Can you run db.version() to verify that.
>
> This messages is not an error. It is just informational.
>
> Have you deleted lots of stuff? You may want to run a repair to
> compact the data. This message is basically saying that there are
> extents which are empty which are being skipped. This means the data
> files are sparse; over time as you add more data they will get filled
> back in and these messages will go away if you don't compact.
>

> On Wed, Jul 6, 2011 at 4:21 AM, 老王 <hi.laow...@gmail.com> wrote:
> > my mongodb is newest stable version. log show lots of msg like bottom:
>
> > Wed Jul  6 16:05:06 [conn67475] info DFM::findAll(): extent 0:2900was> empty, skipping aheaduser_data_analysis.job
> > Wed Jul  6 16:05:07 [conn67475] info DFM::findAll(): extent 0:2900was> empty, skipping aheaduser_data_analysis.job
> > Wed Jul  6 16:05:07 [conn67475] info DFM::findAll(): extent 0:2900was> empty, skipping aheaduser_data_analysis.job
> > Wed Jul  6 16:05:07 [conn67475] info DFM::findAll(): extent 0:2900was> empty, skipping aheaduser_data_analysis.job
>

[Scott Hernandez]

They aren't really related. That lock percentage is reasonably high
but still not the cause of the messages.

It is because you deleted lots of data, and haven't adding as much
back. You can repair/compact the data and those messages will go away.