You are not logged in.
The ideia is beautifull, but i would recomment an fsck to that file from time to time... add it to the options in your script, a simple unmount, fsck and mount should do the trick
Offline
So, you need to run this after every dbase update right? Or what?
Offline
I've done this by adding some lines in rc.local. But some could use cron jobs.
The best way to do this IMHO is imbue this in pacman, have allways the drive unmounted for power failure security and have it checked every 30 times (or less) of pacman use or in pacman fail.
Offline
_Gandalf_: please can you add some rc scripts for this or give a bit more guidance on how and when to use it?
And can anyone explain:
You're missing a major thing here, never use jfs/xfs for pacman database, pacman database is an ext2 partition not xfs/jfs read the topic man...
Why can't you use XFS or JFS? I'm sure I have seen people using the whole of that for var
Offline
Journal space constraints I would think. And because XFS is prohibitively slow with pacman.
Offline
_Gandalf_: please can you add some rc scripts for this or give a bit more guidance on how and when to use it?
And can anyone explain:
You're missing a major thing here, never use jfs/xfs for pacman database, pacman database is an ext2 partition not xfs/jfs read the topic man...
Why can't you use XFS or JFS? I'm sure I have seen people using the whole of that for var
I think he means his script uses a loopback device for the database with ext2 not the others.
Offline
@dtw XFS/JFS are slow for many small size files, ext2/ext3 is much faster that's why pacman database must be ext2/ext3 and not XFS/JFS, it works for sure but slow, and you don't need any rc.d script, just use it once to create the loopback device and fstab entry and that's it, in case you want to remove the loopback device u can use this script again...
[My Blog] | [My Repo] | [My AUR Packages]
Offline
I got an error:
==> ERROR: Could not create image filesystem.
after the creating loopback file part. Obviously I have DEV_BLK_LOOP support in the kernel.
To get something done, a committee should consist of no more than three persons, two of them absent.
--
My Github
Offline
Thanks _Gandalf_ for this great script! The creation of the pacman-drive went flawlessly and pacman seems to be a lot faster now.. Keep up the good work!
Haven't been here in a while. Still rocking Arch.
Offline
I solved my problem. Where the script says:
yes | mkfs.ext2 -O dir_index -b 1024 -i 1024 -m 0 "${pacman_drive}" > /dev/null 2>&1 ||
the yes thingy didn't work at all, therefore the script exited with an error.
I changed to:
mkfs.ext2 -F -O dir_index -b 1024 -i 1024 -m 0 "${pacman_drive}" > /dev/null 2>&1 ||
thus forcing (-F) the creation of the fs without asking, and now it's ok.
Great work Gandalf
To get something done, a committee should consist of no more than three persons, two of them absent.
--
My Github
Offline
Thanks onearm, I'll change it in my SVN
[My Blog] | [My Repo] | [My AUR Packages]
Offline
Errr! dunno if that's specifically pacman-drive related but 1st posting here won't hurt as it *might* be related.
Got a problem with pacman.db : Had 2 freeze last 7 days while using external USB HDD, and 2day I cannot use pacman.
# pacman -Syu
could not extract kdesdk-3.5.4-1/: Input/output error
could not extract kdesdk-3.5.4-1/desc: Input/output error
$ find /var/lib/pacman -name "*kdesdk*"
find: ./lost+found: Permission non accordée
find: ./extra: Erreur d'entrée/sortie
./local/kdesdk-3.5.4-1
find: AVERTISSEMENT: lien direct (hard link) est erroné pour .: cela est peut-être dû à une anomalie dans le pilote du système de fichiersé L'option -noleaf est automatiquement activée. Les résultats antérieurs peuvent avoir échoués à inclure des répertoires qui auraient dû être recherchés.
And then, still less understandable to me : # pacman -Qs kdesk
config: line 6: all directives must belong to a section
Here's what dmesg read about "ext2" :
Oct 17 13:40:42 llewellyn EXT2-fs error (device loop0): ext2_check_page: bad entry in directory #2025: unaligned directory entry - offset=27648, inode=1296125477, rec_len=9541, name_len=10
Oct 17 13:40:42 llewellyn EXT2-fs error (device loop0): ext2_readdir: bad page in #2025
Oct 17 13:40:43 llewellyn EXT2-fs error (device loop0): ext2_readdir: bad page in #2025
Oct 17 13:40:43 llewellyn EXT2-fs error (device loop0): ext2_readdir: bad page in #2025
Oct 17 13:40:43 llewellyn EXT2-fs error (device loop0): ext2_readdir: bad page in #2025
Oct 17 13:40:44 llewellyn EXT2-fs error (device loop0): ext2_readdir: bad page in #2025
Oct 17 13:40:53 llewellyn EXT2-fs error (device loop0): ext2_readdir: bad page in #2025
Oct 17 13:40:53 llewellyn EXT2-fs error (device loop0): ext2_readdir: bad page in #2025
Oct 18 00:03:19 llewellyn EXT2-fs error (device loop0): ext2_check_page: bad entry in directory #2025: unaligned directory entry - offset=27648, inode=1296125477, rec_len=9541, name_len=10
Does someone has an idea about what to do now ? At least some instructions would be much appreciated
Cagnulein > pacman-cage (April 2006) uploaded to your FTP
Seeded last month: Arch 50 gig, derivatives 1 gig
Desktop @3.3GHz 8 gig RAM, linux-ck
laptop #1 Atom 2 gig RAM, Arch linux stock i686 (6H w/ 6yrs old battery ) #2: ARM Tegra K1, 4 gig RAM, ChrOS
Atom Z520 2 gig RAM, OMV (Debian 7) kernel 3.16 bpo on SDHC | PGP Key: 0xFF0157D9
Offline
The file isn't available for download... :twisted:
Offline
gothicknight, we now Gandalf's site is quite broken since quite some time. Why don't you take the file on Cagnulein's FTP while it's there ?
Seeded last month: Arch 50 gig, derivatives 1 gig
Desktop @3.3GHz 8 gig RAM, linux-ck
laptop #1 Atom 2 gig RAM, Arch linux stock i686 (6H w/ 6yrs old battery ) #2: ARM Tegra K1, 4 gig RAM, ChrOS
Atom Z520 2 gig RAM, OMV (Debian 7) kernel 3.16 bpo on SDHC | PGP Key: 0xFF0157D9
Offline
Cagnulein > pacman-cage (April 2006) uploaded to your FTP
Thanks man
Offline
Uploaded to my group server http://gcx.ipg.pt/~gcx07/pacman-cage
Offline
pacman-drive is in archlinux.fr repository:
[archlinuxfr]
Server = http://repo.archlinux.fr/i686
Last edited by wain (2007-07-12 18:07:20)
Offline
Dead?
Offline
Dead?
pacman-cage works great for me with pacman3 and yaourt
pacman-drive should work too.
what do you mean by "dead" ?
Offline
Oh NO NO and NO!!!
P.S: /var/lib/pacman contain also current/extra etc... which can be deleted safely but i suggest not to play with them unless u know what your're doing...
What problems might deleting these (/var/lib/pacman/current , etc) give?
Will the results of pacman -Ss , etc be affected?
March Linux : An Arch Linux "distrolet" that I am trying to develop (March = My Arch!)
Please take a look......:)
Offline
Hello people,
Forgive me for intruding into this conversation, but I do need more information. I've been using Arch since 0.4, and I have never once seen pacman print out a "Please wait" style message, nor have I ever noticed any latency while pacman crunches it's database.
My desktop is a very ordinary P4 Northwood with a Samsung Spinpoint HD (which was designed to be quiet, not fast). I use ReiserFS on all my partitions.
What would I need to change in my setup to make pacman take so much time to read a directory tree?
PS: I will not accept "older hardware" as an answer.
Ciao
Offline
Hello people,
Forgive me for intruding into this conversation, but I do need more information. I've been using Arch since 0.4, and I have never once seen pacman print out a "Please wait" style message, nor have I ever noticed any latency while pacman crunches it's database.
My desktop is a very ordinary P4 Northwood with a Samsung Spinpoint HD (which was designed to be quiet, not fast). I use ReiserFS on all my partitions.
What would I need to change in my setup to make pacman take so much time to read a directory tree?
PS: I will not accept "older hardware" as an answer.
Ciao
Not "old hardware" but "old pacman database". With time, pacman becomes slower due to fragmentation.
2 solutions:
- run pacman-optimize frequently
- user pacman-drive or pacman-cage to mount pacman database in a loopback file
Offline