unofficial linux-ck repo with cpu optimized packages is ONLINE

graysky · 2014-05-31 22:06:21

@keepitsimp - Nope, you just need to sign it:

pacman-key --lsign-key 5EE46C4C

keepitsimpleengineer · 2014-06-01 00:15:42

graysky wrote:

@keepitsimp - Nope, you just need to sign it:
pacman-key --lsign-key 5EE46C4C

Thanks - I must be asleep at then console

dlh · 2014-06-01 20:47:31

I am experiencing such bug with my ssd drive, everything works fine with arch stock kernel.
I am not using bfq for the record just noop on my Crucial M500 120 GB

cze 01 22:26:10 HALF-BLOOD kernel: sd 1:0:0:0: [sdb]  
cze 01 22:26:10 HALF-BLOOD kernel: Result: hostbyte=0x00 driverbyte=0x08
cze 01 22:26:10 HALF-BLOOD kernel: sd 1:0:0:0: [sdb]  
cze 01 22:26:10 HALF-BLOOD kernel: Sense Key : 0xb [current] [descriptor]
cze 01 22:26:10 HALF-BLOOD kernel: Descriptor sense data with sense descriptors (in hex):
cze 01 22:26:10 HALF-BLOOD kernel:         72 0b 00 00 00 00 00 0c 00 0a 80 00 00 00 00 00 
cze 01 22:26:10 HALF-BLOOD kernel:         00 00 00 3f 
cze 01 22:26:10 HALF-BLOOD kernel: sd 1:0:0:0: [sdb]  
cze 01 22:26:10 HALF-BLOOD kernel: ASC=0x0 ASCQ=0x0
cze 01 22:26:10 HALF-BLOOD kernel: sd 1:0:0:0: [sdb] CDB: 
cze 01 22:26:10 HALF-BLOOD kernel: cdb[0]=0x93: 93 08 00 00 00 00 06 a4 57 a0 00 00 00 08 00 00
cze 01 22:26:10 HALF-BLOOD kernel: end_request: I/O error, dev sdb, sector 111433632
cze 01 22:26:10 HALF-BLOOD kernel: EXT4-fs (sdb2): discard request in group:305 block:2548 count:1 failed with -5
cze 01 22:26:10 HALF-BLOOD kernel: ata2: EH complete
cze 01 22:26:18 HALF-BLOOD systemd[1]: Starting Getty on tty2...
cze 01 22:26:18 HALF-BLOOD systemd[1]: Started Getty on tty2.
cze 01 22:26:40 HALF-BLOOD kernel: ata2.00: exception Emask 0x1 SAct 0xc000000 SErr 0x0 action 0x6 frozen
cze 01 22:26:40 HALF-BLOOD kernel: ata2.00: irq_stat 0x40000008
cze 01 22:26:40 HALF-BLOOD kernel: ata2.00: failed command: SEND FPDMA QUEUED
cze 01 22:26:40 HALF-BLOOD kernel: ata2.00: cmd 64/01:d0:00:00:00/00:00:00:00:00/a0 tag 26 ncq 512 out
                                            res 40/00:d4:00:00:00/00:00:00:00:00/a0 Emask 0x1 (device error)
cze 01 22:26:40 HALF-BLOOD kernel: ata2.00: status: { DRDY }
cze 01 22:26:40 HALF-BLOOD kernel: ata2.00: failed command: READ FPDMA QUEUED
cze 01 22:26:40 HALF-BLOOD kernel: ata2.00: cmd 60/08:d8:a8:2c:41/00:00:01:00:00/40 tag 27 ncq 4096 in
                                            res 40/00:00:00:00:00/00:00:00:00:00/00 Emask 0x5 (timeout)
cze 01 22:26:40 HALF-BLOOD kernel: ata2.00: status: { DRDY }
cze 01 22:26:40 HALF-BLOOD kernel: ata2: hard resetting link
cze 01 22:26:41 HALF-BLOOD kernel: ata2: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
cze 01 22:26:41 HALF-BLOOD kernel: ata2.00: supports DRM functions and may not be fully accessible
cze 01 22:26:41 HALF-BLOOD kernel: ata2.00: supports DRM functions and may not be fully accessible
cze 01 22:26:41 HALF-BLOOD kernel: ata2.00: configured for UDMA/133
cze 01 22:26:41 HALF-BLOOD kernel: ata2.00: device reported invalid CHS sector 0
cze 01 22:26:41 HALF-BLOOD kernel: ata2.00: device reported invalid CHS sector 0
cze 01 22:26:41 HALF-BLOOD kernel: sd 1:0:0:0: [sdb]  
cze 01 22:26:41 HALF-BLOOD kernel: Result: hostbyte=0x00 driverbyte=0x08
cze 01 22:26:41 HALF-BLOOD kernel: sd 1:0:0:0: [sdb]  
cze 01 22:26:41 HALF-BLOOD kernel: Sense Key : 0xb [current] [descriptor]
cze 01 22:26:41 HALF-BLOOD kernel: Descriptor sense data with sense descriptors (in hex):
cze 01 22:26:41 HALF-BLOOD kernel:         72 0b 00 00 00 00 00 0c 00 0a 80 00 00 00 00 00 
cze 01 22:26:41 HALF-BLOOD kernel:         00 00 00 00 
cze 01 22:26:41 HALF-BLOOD kernel: sd 1:0:0:0: [sdb]  
cze 01 22:26:41 HALF-BLOOD kernel: ASC=0x0 ASCQ=0x0
cze 01 22:26:41 HALF-BLOOD kernel: sd 1:0:0:0: [sdb] CDB: 
cze 01 22:26:41 HALF-BLOOD kernel: cdb[0]=0x93: 93 08 00 00 00 00 06 a4 30 10 00 00 00 08 00 00
cze 01 22:26:41 HALF-BLOOD kernel: end_request: I/O error, dev sdb, sector 111423504
cze 01 22:26:41 HALF-BLOOD kernel: EXT4-fs (sdb2): discard request in group:305 block:1282 count:1 failed with -5
cze 01 22:26:41 HALF-BLOOD kernel: ata2: EH complete
cze 01 22:26:41 HALF-BLOOD kernel: hda-intel: IRQ timing workaround is activated for card #0. Suggest a bigger bdl_pos_adj.

Last edited by dlh (2014-06-01 20:47:55)

Earnestly · 2014-06-01 23:34:20

A few of those are unsupported NCQ commands, some SSDs don't support NCQ (which is primarily a HDD seeking technique) and can be disabled. Information for how is on the wiki.

Last edited by Earnestly (2014-06-01 23:35:31)

dlh · 2014-06-02 08:15:16

Did you miss everything works fine with arch stock kernel?

dlh · 2014-06-02 18:49:51

Seems like on stock kernel it happens also. Do you know how to turn off only for one disk via grub?

graysky · 2014-06-02 18:55:26

Not sure, dlh. Don't see that on my ssd also using noop.

EDIT: Ah, I see you experience these in the stock kernel or in linux-ck so definitely unrelated. Perhaps hardware is failing? Please open a new thread since not repo related.

barrikad · 2014-06-04 06:34:38

A 1Gb uplink European linux-ck repo mirror may be coming soon.

Last edited by barrikad (2014-06-04 06:47:15)

nTia89 · 2014-06-05 07:29:44

@graysky
thank you for repo-ck, I'm using linux-ck-kx-x86_64 for my opteron and it rocks!

I must report you problem with your server: every ~10MB of downloaded data the connection fall down

graysky · 2014-06-05 08:56:57

That is the quality service godaddy provides. Pacman should resume and it shouldn't happen too often.

If you look at the repo-ck stats, an over-time average of 14 % of all transfers end in the apache code "error 206" which is an I understand it, partial content. It does vary from month-to-month though. This month has been particularily bad at 25 % when comparing transferred bytes which is not just counting bad transfers. The past few months have been better as you can see: Feb for example transferred 636 gigs of data and only 5 % was partial content. April was 9 % of 564 gigs. You can do the math :p

Last edited by graysky (2014-06-05 09:03:37)

barrikad · 2014-06-07 11:14:46

graysky wrote:

...

Contact me when you get there so we can set this up and tweak.

graysky · 2014-06-08 12:05:25

All - Based on a lack of sustained usage of in a number of i686 packages, and since each arch/cpu adds to my build and management time, I have decided to remove support for repo-ck packages in several i686 flavors and one x86_64 flavor shown below. This decision mainly affects i686. I will continue supporting the more or less full line-up in x86_64.

If you are an affected user of repo-ck package denoted by red strike-out text in the list below, please switch to the ck-generic package group (pacman should inform you that ck-xxx conflicts with generic ask if it should uninstall the ck-xxx packages automatically):

 pacman -S ck-generic

List of x86_64 packages:
x86_64-generic
x86_64-atom
x86_64-core2
x86_64-nehalem
x86_64-sandybridge
x86_64-ivybridge
x86_64-haswell
x86_64-p4
x86_64-kx
x86_64-k10
x86_64-barcelona
x86_64-bobcat
x86_64-bulldozer
x86_64-piledriver

List of i686 packages:
i686-generic
i686-atom
i686-core2
i686-p4
i686-pentm
i686-kx
i686-nehalm
i686-sandybridge
i686-ivybridge
i686-barcelona
i686-bobcat
i686-bulldozer
i686-k10
i686-piledriver

Last edited by graysky (2014-06-08 12:21:00)

Alad · 2014-06-09 18:55:05

graysky wrote:

That is the quality service godaddy provides. Pacman should resume and it shouldn't happen too often.
If you look at the repo-ck stats, an over-time average of 14 % of all transfers end in the apache code "error 206" which is an I understand it, partial content. It does vary from month-to-month though. This month has been particularily bad at 25 % when comparing transferred bytes which is not just counting bad transfers. The past few months have been better as you can see: Feb for example transferred 636 gigs of data and only 5 % was partial content. April was 9 % of 564 gigs. You can do the math

I've been having these issues over the last week as well. An easy workaround is to uncomment wget in /etc/pacman.conf, as it automatically resumes downloads (perhaps this may be added to the repo-ck wiki page).

Last edited by Alad (2014-06-09 18:55:58)

Jristz · 2014-06-09 19:24:28

Sometimes I think I'm the only one using i686-atom

Well That could save space on the server right? Cause Maybe keep the previous kernel too could help to those with problems

Alad · 2014-06-10 12:14:34

ck-core2 3.14.6-2-ck works fine, ck-haswell froze at initramfs

Machine:   System: ASUS product: All Series
           Mobo: ASUSTeK model: H81M-PLUS v: Rev X.0x
           Bios: American Megatrends v: 0805 date: 02/20/2014
CPU:       Dual core Intel Celeron G1820 (-MCP-) cache: 2048 KB
           flags: (lm nx sse sse2 sse3 sse4_1 sse4_2 ssse3 vmx) bmips: 10794 
           Clock Speeds: 1: 786 MHz 2: 786 MHz

# gcc -c -Q -march=native --help=target | grep march
  -march=                     		haswell

Last edited by Alad (2014-06-10 12:15:30)

keepitsimpleengineer · 2014-06-10 16:28:08

Jristz wrote:

Sometimes I think I'm the only one using i686-atom

Then I make two...

keepitsimpleengineer · 2014-06-10 16:35:07

graysky wrote:

That is the quality service godaddy provides. Pacman should resume and it shouldn't happen too often.

Twice today. Restart required, handled it without fuss. Less worse than a lot of techno•agony around.

Thanks for all the great work, graysky. Please point your elbow at the ceiling and pat yourself on the back.

graysky · 2014-06-10 19:18:45

You're welcome, enjoy.

masnormen · 2014-06-11 08:37:25

This s**t is getting serious. I should re-run the pacman command like 20 times to download a single linux-ck package! But thanks to Xfer command settings in pacman.conf!

Yeah, but seriously, why is GoDaddy doing this kind of thing?

graysky · 2014-06-11 11:30:21

I think it's because repo-ck moves lots of data and they [godaddy] want me to get sick of them and drop my unlimited service they have me in at an old rate. They have offered to upgrade me to a dedicated server for higher reliability several times now for a large increase in my monthly fee. Each time I remind them that my plan is unlimited and that they can upgrade me without a charge. Bastards. I am working slowly with barrikad to setup a mirror on his service. I have been very busy with work/family shit and traveling though. I am booked this weekend as well but hope to do some testing and setup with him shortly.

Last edited by graysky (2014-06-11 11:33:41)

clfarron4 · 2014-06-11 12:00:48

graysky wrote:

I think it's because repo-ck moves lots of data and they [godaddy] want me to get sick of them and drop my unlimited service they have me in at an old rate. They have offered to upgrade me to a dedicated server for higher reliability several times now for a large increase in my monthly fee. Each time I remind them that my plan is unlimited and that they can upgrade me without a charge. Bastards. I am working slowly with barrikad to setup a mirror on his service. I have been very busy with work/family shit and traveling though. I am booked this weekend as well but hope to do some testing and setup with him shortly.

I'm thinking of setting up a Public Dropbox repository for my packages (and I don't mind the $99, once I actually have the spare cash, hopefully on my birthday in September).

Scimmia · 2014-06-12 03:31:48

I see 3.14.7-1 on the server, but it doesn't seem to be in the database (yet?).

graysky · 2014-06-12 08:59:03

When I do an overnight build, I don't like to blindly index and push to the server. It's indexed and updated now.

barrikad · 2014-06-12 12:10:41

The mirror is set up and configurated. When it get traffic I will be able to tweak more to give you a better experience upgrading to the latest linux-ck.

Please give graysky some slack, the guy has a lot of stuff going on right now.

justinthegeek · 2014-06-12 13:15:31

So I'm using linux-ck on my Core i5 laptop and gcc says I should be using westmere. I'm currently using linux-ck-nehalem because that seems to be the same generation. Is that correct or should I use a different version?

justinthegeek ~ $ gcc -c -Q -march=native --help=target | grep march
  -march=                     		westmere

Thank you graysky for the time and effort you put into this repo!

Arch Linux

#1851 2014-05-31 22:06:21

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1852 2014-06-01 00:15:42

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1853 2014-06-01 20:47:31

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1854 2014-06-01 23:34:20

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1855 2014-06-02 08:15:16

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1856 2014-06-02 18:49:51

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1857 2014-06-02 18:55:26

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1858 2014-06-04 06:34:38

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1859 2014-06-05 07:29:44

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1860 2014-06-05 08:56:57

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1861 2014-06-07 11:14:46

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1862 2014-06-08 12:05:25

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1863 2014-06-09 18:55:05

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1864 2014-06-09 19:24:28

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1865 2014-06-10 12:14:34

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1866 2014-06-10 16:28:08

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1867 2014-06-10 16:35:07

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1868 2014-06-10 19:18:45

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1869 2014-06-11 08:37:25

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1870 2014-06-11 11:30:21

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1871 2014-06-11 12:00:48

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1872 2014-06-12 03:31:48

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1873 2014-06-12 08:59:03

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1874 2014-06-12 12:10:41

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

#1875 2014-06-12 13:15:31

Re: unofficial linux-ck repo with cpu optimized packages is ONLINE

Board footer