Hi all, Thanks you for your information. It is a clear now that there is a big difference between the index files from yours and ours. Geert Jan de Groot <GeertJan.deGroot@ripe.net> writes: * * We're currently trying to break out own test-database (controlled * environment) too, though have not succeeded. I tried today to index the database with netdbm and cleandb on our SUN machines. Both programs successfully did the job :-(. Currently, I am trying the same on a BSDI machine. Since we are still not able to reproduce the problems we would like to hear answers on the previously posted list from everybody experiencing the same problems: In order for us to be able to reproduce this problem, we would like people that encounter it send reports to ripe-dbm@ripe.net with the following information: 1. ls -ls of the files involved 2. Do you 'clean' the database frequently? 3. perl -v? 4. ls -ls of the bin directory of the database 5. egrep $Id of the same bin directory 6. You OS & version number, patches? 7. If you know, what dbm you are using with perl If you don't know could you try to find out which one you use ... *New question:* 8. the contents of the script you use for indexing the database Then, I would like to ask everybody who experiences these problems if time permits: To try out: - another perl version if available - if you use 'netdbm' for indexing, to try out 'cleandb' or the other way around. You will certainly hear more about this, thanks for your help so far, David Kessens RIPE NCC --- * * One point people missed is that I want ls -ls * (mind the -ls at the end); this shows how big the index * files really are * * On Mon, 25 Sep 1995 11:17:02 +0100 (MET) Willi Huber wrote: * > > 1. ls -ls of the files involved * > * > -rw-r--r-- 1 huber 7143967 Sep 24 09:03 ripe.db.in * > -rw-r--r-- 1 huber 222257152 Sep 24 11:37 ripe.db.in.cl.dir * > -rw-r--r-- 1 huber 1728317440 Sep 24 11:37 ripe.db.in.cl.pag * * We have: * 6992 -rw-r--r-- 1 auto-dbm 7148193 Sep 25 11:14 ripe.db.in * 19 -rw-rw-rw- 1 auto-dbm 19185 Sep 25 10:25 ripe.db.in.0.0.0.0.0 * 64 -rw-r--r-- 1 auto-dbm 65046 Sep 24 02:35 ripe.db.in.192.0.0.0.8 * 4 -rw-r--r-- 1 auto-dbm 3350 Sep 24 02:32 ripe.db.in.192.109.0.0.16 * 3 -rw-r--r-- 1 auto-dbm 2314 Sep 24 02:08 ripe.db.in.192.164.0.0.16 * 4 -rw-r--r-- 1 auto-dbm 3252 Sep 24 02:18 ripe.db.in.192.165.0.0.16 * 7 -rw-r--r-- 1 auto-dbm 7156 Sep 24 02:34 ripe.db.in.193.0.0.0.8 * 4 -rw-r--r-- 1 auto-dbm 3391 Sep 24 02:33 ripe.db.in.193.100.0.0.15 * 3 -rw-r--r-- 1 auto-dbm 2314 Sep 24 02:34 ripe.db.in.193.102.0.0.16 * ... * 96 -rw-r--r-- 1 auto-dbm 475136 Sep 25 10:50 ripe.db.in.cl.dir * 6824 -rw-r--r-- 1 auto-dbm 1728316416 Sep 25 11:14 ripe.db.in.cl.pag * 4 -rw-r--r-- 1 auto-dbm 4096 Sep 24 02:22 ripe.db.in.dir * 3024 -rw-r--r-- 1 auto-dbm 8349696 Sep 25 11:14 ripe.db.in.pag * * ... as see, the dbm-files are not as big as they look at first. * For people unfamiliar with this, UNIX files can have 'holes' in them. * Blocks not written into are not allocated. A file made by * open() * fseek(16M) * write(1 byte) * close() * will not have a size of 16 Mbyte. * For this reason, you cannot copy the index files using cp * (I think you all know, but just want to check) * * Willi et all, * Can you send us ls -ls again? * * Geert Jan * (still looking for the red wire in the database-timebomb)