Monitoring the Syslog

[BLKMGK]

Tailing the syslog with tail -f /var/log/syslog I see some errors pop up occasionally. Overallnothing too alarming except something about an inode being hosed with a fsck? commetn afterwards. I'm no Linux Wiz but it looks like there might be some corruption and that the OS wants a file check? This brings up a question, if something gets a little borked with Reiser is there a way to correct it within unRAID?

 

Here's what I'm seeing presently, I've not got the fsck stuff displayed at the moment..

 

Mar 24 12:22:20 BEAST smbd[2133]:  sys_path_to_bdev() failed for path [.]!

Mar 24 12:22:20 BEAST smbd[2133]: [2006/03/24 12:22:20, 0] lib/sysquotas.c:sys_get_quota(386)

Mar 24 12:22:20 BEAST smbd[2133]:  sys_path_to_bdev() failed for path [.]!

 

I see the failed comments pretty often, sometimes all I have to do is refresh the WEB interface to make it appear. System appears to be running smothly though.

[lovingHDTV]

My syslog if full of these also.  I was hoping the new code would fix it.

 

 

[BLKMGK]

My syslog if full of these also.  I was hoping the new code would fix it.

 

 

A okay, so it was there before? I feel better then actually I had only monitored the logs for a short while before updating to the new code so I was under the impresion that this was new since I hadn't seen it before. naturally since my previous post the fsck? thing hasn't shown back up. Will post when I spot it but the other stuff seems pretty benign since my system ran just fine before... I've yet to see one of the dreaded DMA errors BTW but I had pretty reliable performance before too.

 

I'm presently ripping and moving alot of code to the array so at least one disk is seeing a good bit of activity. Multiple machines writing at once or trying to ripacross the network is a pretty bad idea -the array cannot hande it. Playback from it appears to be just fine though!

[TCIII]

BLKMGK,

 

I have no problem ripping my DVDs to the unRaid server in the same amount of time that it takes to rip locally.

 

I can usually rip a two hour DVD to either the local HD or the unRaid server in 10 to 15 minutes.

 

I do not do ISO. I just rip the main title and the Dolby AC3 sound track without compression. No extras, no menus.

 

To say that ripping to the server is a "bad idea" might just be isolated to your setup.

 

Just a thought.

 

Regards,

TCIII

[BLKMGK]

BLKMGK,

 

I have no problem ripping my DVDs to the unRaid server in the same amount of time that it takes to rip locally.

 

I can usually rip a two hour DVD to either the local HD or the unRaid server in 10 to 15 minutes.

 

I do not do ISO. I just rip the main title and the Dolby AC3 sound track without compression. No extras, no menus.

 

To say that ripping to the server is a "bad idea" might just be isolated to your setup.

 

Just a thought.

 

Regards,

TCIII

 

Well, I have one VERY fast machine with GigE on the same switch as my unRAID. My first attempt to rip with that machine took 6.5 HOURS for a single DVD that only compressed the extras. Another machine that started ripping just after the first also took 6 hours but is a MUCH slower CPU. Displayed bitrate for one was about 900K and the other maybe 1500K. If I rip local with either machine or even a 3rd still slower machine I see bitrates double and triple what I'd seen before and can complete a DVD in 11-25mins. That's ISO or folder format - I'm now doing just folders BTW but keeping extras, menus, etc.

 

I will admit that since that first attempt I've not tried to rip across the network since it was such a disaster and that perhaps multiple machines working was the problem but 6 hours compared to say 30mins is a pretty drastic difference! CPUs weren't maxxed and network traffic was low while that was going on, I don't think it was an issue with my various machines. Copying the resulting multi-gig files to the server from local now takes maybe 6-7mins unless I try to do it with multiple machines, at that point performance goes WAY down but nowhere near the hours long mark. <shrug>

 

Hope that provides some insight anyway!

 

P.S. I said 2-3times performance ripping locally. I just noticed that the bitrate for my second fastest machine is reporting nearly 8 meg a second so umm maybe a "little" more than 3 times Both mytwo fastest machines have GigE onboard, one on th emotherboard (Nvidia) and another a Hawk add-on card. switch is an SMC unit - a CG58

[limetech]

We're looking at performance issues now - trying to make writes faster & let reads have priority over writes enough to prevent lost video frames.  Current s/w write performance peaks somewhere around 11MB/sec. (11 mega-Bytes per second).  There are both network issues and disk issues to address.

[BLKMGK]

We're looking at performance issues now - trying to make writes faster & let reads have priority over writes enough to prevent lost video frames.  Current s/w write performance peaks somewhere around 11MB/sec. (11 mega-Bytes per second).  There are both network issues and disk issues to address.

 

I understand, I honestly didn't expect the newest software to solve things as you'd stated that wasn't the focus. So long as I don't multitask the server it seems to work better. I just tried ripping across the network again and was seeing over 2meg per second and that usually ramps up as I read towards the edge of the disk. Unfortunatly at 6% the software froze, below is some of what I see in the logs. I went back to a local rip and it's flying - at 46% I'm seeing nearly 11meg per second

 

Mar 24 19:13:05 BEAST smbd[1079]: [2006/03/24 19:13:05, 0] smbd/service.c:make_connection(798)

Mar 24 19:13:05 BEAST smbd[1079]:  192.168.1.250 (192.168.1.250) couldn't find service disk

 

Not sure if that helps at all. The software itself didn't crash and I was able to recover and start again local <shrug> Wonder if it doesn't like UNC naming? TyTools doesn't I don't think and neither does PSP Video9 as I recall. The software I'm using right now is DVD Shrink FWIW.

[BLKMGK]

Ah Ha!!!!

 

One of the directories I've been unable to delete is the cause of the Fsck? errors in the log! When I attempt to delete it from Windows it reports that the directory isn't empty but nopthing I try in Windows displays anything. I've not tried navigating to it in Linux. I strongly suspect that at some point a filename with High ASCII was in the directory but I'm not positive. I've had filenames with Kanji and other languages trip me up, I think I've gotten rid of them all though.

 

Below is what I've been spotting in the log and when I access this directory or attempt to delete it the log goes a bit nutz

 

Mar 25 02:29:48 BEAST kernel: md(9,5):vs-5150: search_by_key: invalid format found in block 40594524. Fsck?

Mar 25 02:29:48 BEAST kernel: md(9,5):vs-13070: reiserfs_read_inode2: i/o failure occurred trying to find stat data of [34613 34833 0x0 SD]

Mar 25 02:29:48 BEAST kernel: is_tree_node: node level 1709 does not match to the expected one 1

Mar 25 02:29:48 BEAST kernel: md(9,5):vs-5150: search_by_key: invalid format found in block 40595373. Fsck?

Mar 25 02:29:48 BEAST kernel: md(9,5):vs-13070: reiserfs_read_inode2: i/o failure occurred trying to find stat data of [34613 34834 0x0 SD]

 

 

Umm, how best to fix this? Kill the directory in Linux or try to fixup the disk? This is my MP3 disk so a "bad thing" would be VERY bad indeed