1-3.1 seem correct > 3.2) In case of RAW storage-method there is no file system and I have to > calculate offset for every new table. How can I do that? TableConvert tells you how many bytes were written (and if you are converting to files, you will see this as a file size). I'd round this number up to the nearest multiply of 4096 bytes and use this as an offset for the next one.
I don't know how much performance penalty is imposed by having tables in files instead of the raw device. I hope not so much and I'd leave it in files at least during experimenting. On 18.1.2016 17:12, Вадим Яницкий wrote: > Thanks for explanations! > > Though I'd better keep the original files in case you find out something > is screwed up. > > > It may sound sad, but I have no more than 2 Tb of space. > I have no even direct HDD access because my server is virtualized and running > on the Hyper-V (it's not my own). > > If I understood correctly, I need to do the following actions: > 1) Prepare 2 Tb HDD changing the ownership using `chown`. > 1.1) In case of file system storage-method format it to ext4, for example. > 2) Download the first part (a51_table_100.dlt) to the home folder > of my "buffer-hdd" (60 Gb). > 3.1) In case of file system storage-method I should just run: > ./TableConvert di (/dev/sda)/home/$file.dlt \ > (/dev/sdb)/$file.ins:0 (/dev/sda)/home/kraken/indexes/$file.idx > on every downloaded file. And in this case the offset param will always = 0. > 3.2) In case of RAW storage-method there is no file system and I have to > calculate offset for every new table. How can I do that? > > Is this sequence of actionsright? > > With best regards, > Vadim Yanitsky. > > 2016-01-17 17:49 GMT+06:00 Jan Hrach <[email protected] > <mailto:[email protected]>>: > > > I'm new in A5/1 world and I have read that 2 Tb is not enough for table > conversation. > > It is (you can convert it one-by-one, requiring only 40 additional GB). > Though I'd better keep the original files in case you find out something is > screwed up. > See https://brmlab.cz/project/gsm/deka/deka-admin > > > Can anyone explain me the Behemoth.py configuration format? What means > this section: "#Tables: dev id(advance) offset"? > The hash-mark lines are the comments. > I think you want to tell it somehow how many tables should be installed > on which device, but that's not the case with one-by-one installation. > > If you have it on filesystem as files, I'd just run > ./TableConvert di $file.dlt $file.ins:0 $file.idx > on every file. You can do the block device installation later with dd > seek=X. > > > On 17.1.2016 10:23, Вадим Яницкий wrote: > > Hi folks! > > > > I'm new in A5/1 world and I have read that 2 Tb is not enough for table > conversation. > > > > But I have a server with 2 Tb and 60 Gb of additional space only. > > My question is: is it possible to convert the tables one by one? I.e. > can I download first table (about 41 Gb) on my 60 Gb HDD and convert it, > download next one and convert with offset ...? > > > > And one more question. Can anyone explain me the Behemoth.py > configuration format? What means this section: "#Tables: dev id(advance) > offset"? Maybe it is an answer to my previous question. :) > > > > With best regards, > > Vadim Yanitskiy. > > > > > > _______________________________________________ > > A51 mailing list > > [email protected] <mailto:[email protected]> > > https://lists.srlabs.de/cgi-bin/mailman/listinfo/a51 > > > > -- > Jan Hrach | http://jenda.hrach.eu/ > GPG CD98 5440 4372 0C6D 164D A24D F019 2F8E 6527 282E > > > > > _______________________________________________ > A51 mailing list > [email protected] > https://lists.srlabs.de/cgi-bin/mailman/listinfo/a51 > -- Jan Hrach | http://jenda.hrach.eu/ GPG CD98 5440 4372 0C6D 164D A24D F019 2F8E 6527 282E _______________________________________________ A51 mailing list [email protected] https://lists.srlabs.de/cgi-bin/mailman/listinfo/a51
