Slow network, No LAGG nics, Erratic speed and also lost network for a while.

[Alucard1972]

Sorry to post new thread but searching posts since fri and freenas 8 posts seem slim.
Alot of similar issues have been delt with but im yet to find my solutions.
System specs, 720w gigabyte psu, gigabyte ga-ma785gt-ud3h (bios 8)RTL8111c onboard lan, amd 265 dual core 3.3ghz (i think), 12 gig ddr3 1333, 6*2tb hdds raidz2, zfs single pool, tplink TG-3269 gigabit card Realtek RTL8169SC, kingston 4 gig usb thumbdrive as OS (psu, main and 2 hdds are old all others brand new)
Slow and erratic gig lan speeds to freenas 8 box, approx 30mb BUT (similar post) i can copy two files at once and get 60mb+, almost as if each connection is limited.
Cpu is at 100% when copying, swap is 12g free/unused, memory is 1 gig free.
Alot of network destination not found errors when copying to the box, mid copys anytime really.
Both nics do not appear in gui lagg setup, dos interface isnt much better.
Tried 1 card at once, disabled onboard in bios, auto config, minimal config, dhcp, not, ip4, gui network nic setup, dns.
Im really at a loss troubleshooting this it could all be 1 setting i missed or just the way freenas 8 is atm thats why im asking.
The freenas box is on gigabit network with windows 7 and vista machines, also mvix v8 reads from it too.
So i have installed, set lan up with gateway, ip4 and dns, named both connections too, switched on cifs and configured cifs workgroup name, send file 2 large rw support maybe a few other boxs ticked im not sure anymore.
Other than the network being slow erratic and dropping out i was also unable to find which hdd failed when i removed two as a test what tells you which drive? the sata port, i couldnt see anywhere the problem(removed drive) only system degraded in storage.
Thank you in advance and i hope this may help others.

 

[matthewowen01]

if you have a failed drive, the data on that drive must be recreated for each read. this may be the issue with your speed, you said the cpu is tapped out. try tracking down the failed disk and replace it. fix the known issue first.

 

[Alucard1972]

Sorry but the failed drive was a test, id deliberately removed the drives to see what happens when it does die. the gui only said degraded not ad5 or sata port3 was damaged or even samsung or hitachi drive, where does it inform you of which drive needs attention?
Some progress id made was i added an asus gig lan card and then disabled onboard so only asus and tplink cards were active then in the dos screen it allowed me to make a link aggregation of load balance, the gui also showed it as built but no physical nics still.
I didnt care because i got it working, and as i suspected it made no difference still below avg transfer speed.
If i need to provide more info tell me what to do and ill grab it.
Have i screwed up a windows share setting or something is it just bugs in freenas 8?

 

[matthewowen01]

Degraded means one of the drives is non functional in the pool. get that fixed first. where it says 'degraded' look to the far right there is a button that has a magnify glass on it, click it and it should give you a detailed view of your pool. that should tell you which disk you need to attend too.

 

[Alucard1972]

Excellent, that is the info i was looking for, the magnifying glass at the end that shows all of the individual disks, although i think that still showed all 6 disks as normal attached (when only 4 were) i think that's why i gave up then just reconnected them.
Copied more data across earlier same slow erratic speed, i've adjusted my hdds spindown time from 5 to 20 adjusted acoustic from disabled to medium, enabled smart, set transfer mode to sata 300 and power to 127 on all 6*2tb drives.
test #18225 copied data from freenas to ssd on second pc, 6.26 gig of data approx 180mb files (2 seasons 30 rock with .jpg and .xml's) 7mins 12 seconds approx 22% of gig lan usage, avg 30mbs transfer speed, freenas reporting charts= cpu minimal used, swap not used, ram 10gig free 2 gig used.
"Compared to other networking protocols, CIFS is not fast" this is about the closest answer i have found in the forums to my networking speed although it doesnt explain the erratic variations in speed.
Also just tried aio set to 8192 instead of 1 no difference after the restart.
Seems nothing i do effects this network behaviour besides copying two things at once equaling the speed 1 copy should get, so maybe shoving 6 drives, $150 of ddr3 ram and a new 3.3ghz dual core, with 2 gig nics in load level mode are a waste of time, although i hopefully get the data security and disk redundancy even if it takes three days to copy it off.

 

[matthewowen01]

so your raidz is not reposting 'online' right? it's no longer 'degraded'?


set aio to 1, there is a reason it's 1, because anything that or larger gets aio applied to it, by setting it to 8192, you are limiting it's effectiveness.

cifs is bad for small files. don't test with small files, get a very large file and copy that.

 

[Alucard1972]

Ok, the networking might be an oversight my end tested between ssd's on my network and still only got 30mbs, so ill fix that first.
Back to the disks!!!, well bugger me i had a disk fail already yay for disk fails!!!! so the only reason i knew was flicked telle over to the freenas dos screen and it had written in white writing couldnt connect to ada01 or something like that, changed to web browser window and sure enough DEGRADED!!!.
I tried to see which sata port or any info as to which drive it was but no luck ada 0-5 are all the same under the magnifying glass, i saw nothing on the web browser alerting me to which disk had died, no samsung, western digital, sata port 01, 3, nothing.
The only way i knew is that the dos screen said ada 1, so rebooting through the dos screen i used pause break and scrolled up to the ada and saw it was my older seagate 2tb that was on ada1, so i shut it down removed that drive replaced it with a brand new western digital and rebooted machine.
Dos screen was perfect so flicked to web browser, into disks clicked replace on ada1 waited for a sec then all seemed well and i left it run for the whole day.
I just got home thinking all would be online but cpu has been running at 50% all day and system is still degraded, dos screen was fine. What gives!! isnt this thing supposed to repair itself, did i miss the i dont know crap button or somethnig?
Since getting home i have shut down and restarted the machine incase that is what i didnt do after REPLACING DISK in web browser, or maybe i havent waited long enough for it to repair.
This program really needs to be a bit more user friendly alerting, reporting and just letting you know whats its doing, like "hey linux noob im currently rebuilding the screwed disk i told you to replace" or "i was just running cpu at 100% for three hours cause i wanted to know if you were watching"
That was a joke too incase the free bsd police are reading this, freenas is a great system if you know your way around and follow the rules.
But i dont know my way around and have 5 tb of data on a degraded raidz2 and im a bit stressed because it isnt as simple as i thought.
Any advice would be appreciated, ill have a scab through the forum again for how to identify and replace a damaged drive.
Thank all who asist in advance.

 

[matthewowen01]

Ok, the networking might be an oversight my end tested between ssd's on my network and still only got 30mbs, so ill fix that first.
Back to the disks!!!, well bugger me i had a disk fail already yay for disk fails!!!! so the only reason i knew was flicked telle over to the freenas dos screen and it had written in white writing couldnt connect to ada01 or something like that, changed to web browser window and sure enough DEGRADED!!!.
I tried to see which sata port or any info as to which drive it was but no luck ada 0-5 are all the same under the magnifying glass, i saw nothing on the web browser alerting me to which disk had died, no samsung, western digital, sata port 01, 3, nothing.
The only way i knew is that the dos screen said ada 1, so rebooting through the dos screen i used pause break and scrolled up to the ada and saw it was my older seagate 2tb that was on ada1, so i shut it down removed that drive replaced it with a brand new western digital and rebooted machine.
Dos screen was perfect so flicked to web browser, into disks clicked replace on ada1 waited for a sec then all seemed well and i left it run for the whole day.
I just got home thinking all would be online but cpu has been running at 50% all day and system is still degraded, dos screen was fine. What gives!! isnt this thing supposed to repair itself, did i miss the i dont know crap button or somethnig?
Since getting home i have shut down and restarted the machine incase that is what i didnt do after REPLACING DISK in web browser, or maybe i havent waited long enough for it to repair.
This program really needs to be a bit more user friendly alerting, reporting and just letting you know whats its doing, like "hey linux noob im currently rebuilding the screwed disk i told you to replace" or "i was just running cpu at 100% for three hours cause i wanted to know if you were watching"
That was a joke too incase the free bsd police are reading this, freenas is a great system if you know your way around and follow the rules.
But i dont know my way around and have 5 tb of data on a degraded raidz2 and im a bit stressed because it isnt as simple as i thought.
Any advice would be appreciated, ill have a scab through the forum again for how to identify and replace a damaged drive.
Thank all who asist in advance.

that stress you feel, that's why i don't run raidz anymore ;) and remember, you still have a disk to loose and you'll still be good. Imagen having 8 TB on the line with no disks left, and the UPS tracking says 'train derailed...' true story.


ok, can you ssh into the box and run the command 'zpool status' let me know what it outputs. it's probable your system is still re silvering the data to the new drive. it can take a while depending on how fast your cpu is and how much data there is to replicate.

 

[Alucard1972]

Thank you for your words of support, i dont know how to ssh so i used the dos screen and shell which i guessed was similar and sure enough zpool status gave me results.
Ok it said: scrub completed in 6hrs xmins but system was/is still degraded with that disk ada1, it says replacing state-degraded read-0 write-0 cksum-0 was /dev/gpt/ada1/old.
So i thought maybe the wd 2tb i replaced it with had been formatted or something so i shut it down and threw a new seagate in (same as degraded drive was) and rebooted, went into web gui and replaced disk like last time then opened shell and typed zpool status, now the new disk has its own line of description saying unavail (as wd said too) but this time it says under write 177.
Also at bottom it says errors: No known data errors.
Just a thought, with freenas 7 before you could do anything you had to ADD the disk in the web gui and then formatted it or whatever, have i been that retarded and need to add my new drive to web gui before i tell replace, i say this because in the drop down box there wasnt any choices of replacements, i just assumed it would know id replaced it by clicking REPLACE. ill go have a scab and see if it makes sense.
Thanks