Software :: Device-mapper-multipath - Slow Failover Period
May 10, 2010
I have a linux machine, attached to a fibre channel SAN.
We're in the testing phases and we're attempting to get all of the bugs worked out before this goes live.
If i have my host streaming data to the storage device on the san (or from the device on the san) and simulate a path failure (by shutting down one of the host's ports on the FC switch), multipath does not pick up on another path until about 45 seconds have passed.
I can verify this by watching the statistic graph (which updates once per second) on the storage system.
I see iops running along rather nicely, and then they drop to 0 for 45 seconds, then pick right up to normal again.
This is a RehHat EL 5.5 system, with qlogic HBA's.
Am i being too picky? I'd expect multipath to recover in under 30 seconds, so as to not alarm host applications running on the linux host... 45 seconds seems like a long time to wait for a disk operation to complete.
Any tips on tuning Multipath, or the qlogic card? As it is, i've got the following options in my modprobe.conf.
alias scsi_hostadapter2 qla2xxx
Code:
My multipath.conf looks like:
Code:
View 3 Replies
ADVERTISEMENT
Aug 13, 2009
I am trying to setup multipath with failover policy on openSuSE 11. I have two qla2xxx HBA's installed and they appear to be working. Here is the output of "multipath -l" command
[Code]....
While testing, I pulled one of the two connection to SAN, and the connection failed over to second HBA connection to SAN. When I plug the cable back in, it does not fall back to original connection... It stays in failed state. Also, I noticed that failed disk (sdd disk) comes back as (sdg disk), which is probably why connection does not fall back to original HBA. But, when I run "/sbin/service multipathd restart" sdg disk shows as as enabled in multipath -l...
View 9 Replies
View Related
May 15, 2011
Recently, I created a device sc0 through device mapper. The divice could be found in /dev/mapper/sc0. My problem is that the device doesn't exist in /dev/partitions which will block my following test.BTW, I found dm-0 in /dev/partitions. Is it the same as /dev/mapper/sc0? But the device /dev/dm-0 doesn't exist!
View 2 Replies
View Related
Jun 23, 2011
Dell PowerEdge R610 with QLogic HBA card connected to HP EVA6000 SAN through a pair of switches.When I present space to the server, device mapper doesn't see it unless I restart the multipathd service. Same thing with unpresenting - the /dev/mapper/mpatha does not go away unless I restart multipathd.Here's what it looks like in 'multipath -ll' before and after I unpresent the LUN (without restarting multipathd)...
== BEFORE ==
# multipath -ll
mpathb (3600508b40010716b0000900001a40000) dm-8 HP,HSV200
[code]....
View 3 Replies
View Related
Feb 14, 2010
What is the current state of NFS failover (i.e. setting up two server with shared storage, with automatic failover if one fails?) I've seen a cookbook, but no details that would let me assess how well it works. There are lots of complex issues with data consistency, but the detailed information on that is years old. Our needs are fairly simple: 2 servers, a shared array, and I'm reasonably sure that we don't use locking. However we'd like failover to work reliably without loss of data.
View 3 Replies
View Related
Feb 28, 2011
I'm facing a problem on Boot on San installation, Os is installed, now I'm trying to define mapper device
for others lun/disk with alias for multipath references
Code: # pvcreate /dev/dm-14 /dev/mapper/disk_bin_rac
Device /dev/dm-14 created
[code]....
View 2 Replies
View Related
Mar 3, 2010
Slackware 13 Kernel 2.6.29.6 have three hard drives. Root is on own drive sda1. sdb and sdc are raid via mdadm with two partitions. one for /home raid0 md0 one for swap raid1 md1. md0 is encrypted vi cryptsetup. md1 is encrypted vi fstab. everything boots fine and is accessible. However, /dev/mapper/* shows sda1 as block device connected to the raid md0. swap crypted device is correct in /dev/mapper/*. fstab is set correctly. problem seems to be with initrd. I would like the correct device in /dev/mapper so that I can access drive info; size, available space, etc. now info shows only sda1 info
View 1 Replies
View Related
Mar 5, 2009
I am trying to prevent the dm_mod and friends from loading during install.I am installing by PXE boot with NFS and kickstart files over http. This works fine for other machines.Ok, I have a Dell PowerEdge 2900 III with 8gig ram, PERC 6/i raid controller.The raid has 8 1TB drives in it, so the total size of the device is >2TB.There is also an SATA drive, which is NOT on the raid controller.
The problem is, I want to install CentOS on the SATA drive. The SATA drive comes up as a device-mapper drive, with a big crazy device name.This is ok, but the problem is, when the system goes to boot, it just says 'missing operating system'.I can't boot from the raid because it is large enough to require a GPT partition, and CentOS says it can't boot from a GPT partition.The SATA drive would work FINE if I could just prevent the damn device mapper from loading, so that I could install on the SATA drive in ordinary SATA mode.
I have tried re-squashing the stage2.img with /etc/modprobe.conf with alias dm_mod off and such, no luck.I also tried adding an /etc/modprobe.d/blacklist file, that did not work either.I also tried putting blacklist= dm_mod in the kickstart file, but that seems to be a fedora-only option and it kills my install. any ideas on preventing the dm_mod et. al. from loading at install time??
View 1 Replies
View Related
Feb 26, 2010
Greetings Fellow Knights Of The Penguin Clan....!I am having issues with device mapper seeing a 275GB Raid5 Lun from my SAN storage.I'm using IBM 2145 San Volume Controller.I am able to see a 40 GB Raid 10 device though..
########################################################################
fdisk -l |grep 'dm-*'
Disk /dev/dm-6: 42.9 GB, 42949672960 bytes
[code]....
View 22 Replies
View Related
Sep 8, 2009
Trying to do a yum update to get everything to latest, towards the end it says this:
[Code]...
how to get around this? I tried yum clean all, then yum update, again but it did the same. I had other deps missing on other servers but yum clean all fixed them -- can't find anyone else who's had this specific issue either, nor an rpm called 'device-mapper-event' or the other things that are missing - am kinda stuck!
View 8 Replies
View Related
Jun 13, 2010
I'm having some difficulities with truecrypt and cannot figure out what's wrong since this is my first attempt to create hidden encrypted partition. I'm using ubuntu 9.10 64b with latest 6.3a truecrypt. After using truecrypt gui (just command truecrypt) I've tried: create volume-->create volumne within a partition/drive-->hidden trucrypt volume --> select some /dev/sdX by my choice --> select aes & sha512 --> password --> and format it.
when formating finished message "wrong ss, swith or wrong superblock of /dev/loop0" appeared. I've got no idea what could I do with it neither know whats program trying to do! Ok formating's done, then if you're trying to mount that partition then I would expect prompt for password, but that did not happen ...
EDIT: sudo mount /dev/loop0 /media/tmp/ also complaint about wrong superblock, but that does not ring any bell since i've got no idea what /dev/loop0 is or is for...
EDIT2: I've tried another approach also:
sudo truecrypt -t -c
sudo truecrypt -t /dev/sda3 --filesystem=none
but the first command finished creating 100G drive in 1 sec (nonsense) and latter one prints:
Error: device-mapper: reload ioctl failed: Invalid argument
Command failed
View 3 Replies
View Related
Jul 8, 2011
Anyone know how device-mapper works, or can someone point me to any docco (other than the man page)?
Couple years ago I wrote a TrueCrypt wrapper script for use on Puppy Linux, and am reworking it for Ubuntu. Puppy sometimes didn't have the 'dmsetup' device-mapper utility available, but TrueCrypt still works without this when the '-m nokernelcrypto' option is used.
My script notices when dmsetup is available and skips this option. When the TC volume gets mounted (somehow using dmsetup), the Desktop icon is labled with the leaf name of the /media/mount_point -- so far, so good.
However, if I force the script to add the no-kernel-crypto option, and the TC volume is mounted without using dmsetup, the Desktop icon ls labled with the size of the volume. Say what?
Running df shows the difference between the two mounts is whether /dev/mapper is used. So *how is* dmsetup making this happen? The manpage isn't much help. Is it just the 'rename' sub-command? I am unable to find any details about this mechanism.
Code:
View 1 Replies
View Related
Apr 2, 2009
I run many 5.2 virtual machines inside of VMWare ESX3.5 I've updated a couple of test VMs from 5.2 to 5.3. The thing that stands out is that during boot time, the sequence gets to this stage:
device-mapper: dm-raid45: intialized v0.2429 Waiting for driver initialization.
Here it takes about 3 times longer than the previous Centos 5.2 (about 9 seconds instead of 3). Whereas inside a physical box the wait in 5.3 is the same as it was in 5.2
View 4 Replies
View Related
Nov 24, 2009
I'm running Fedora 12 on an Acer Aspire One with KDE. The built-in wireless works without a problem EVERYWHERE except at work. It's a shared wireless connection using WPA/WPA2 security.I can connect but it's "hit-and-miss" if I can do anything with it. Somedays it works well for a few minutes then has a period of inactivity - then slow - randomly receives...It worked flawlessly when the computer was running Windows7 no one else has connectivity issues but I'm the only one using Linux.
View 2 Replies
View Related
Sep 10, 2009
I've read hundreds of threads on this problem but they seem to be intermittent and the resolution has been using a different kernel version.
I'm using the latest kernel but I still have this problem
Running Logical Volume Manager gui and attempting to extend the volume results in the same error.
View 3 Replies
View Related
Nov 5, 2010
Installed Maverick last night... the system has been working fine all day... then on a reboot tonight got the following.
RAMDISK: gzip image found at block 0
usb 1-2: new high speed usb device using ehci_hcd and address 2
uncompression error
VFS: Cannot open root device "mapper/mylaptop-root" or unknown-block(0,0)
[code]....
Then it gives the Kernel Panic
Kernel Panic - not syncing: VFS: Unable to mount root fs on unknown-block(0,0)
Pid: 1, comm: swapper not tainted 2.6.35-22-generic #35-Ubuntu
Then a call trace...This is a brand new install of Maverick I copied my files onto a USB and did a fresh install on the whole drive using the alternate CD (the desktop and netbook editions both failed to read on my system) was previously using Karmic with no issues. I tried to e2fsck the dev/sda1 from the CD in "repair broken system" mode but the return was "clean" I read that as the file system being intact but this is an area that I have no real knowledge in.
View 1 Replies
View Related
Jul 8, 2010
I have installed ubuntu 9.04 version. But now its too slow to copy files to any devices or even to other drives.
View 3 Replies
View Related
Jul 16, 2010
Recently my Ubuntu 10.04 was booting slow and to figure out what was going wrong, I booted Ubuntu in text mode. There I found it was hanging on for 5-6 sec showing "unable to enumerate usb device to Port 1". I know it has something to do with Port1 / usb device, but could not understand and solve it.
View 3 Replies
View Related
Sep 9, 2010
Recently did an apt update and upgrade to my CLI only Lenny server. Upon reboot I get an "ATA softreset failed (device not ready)" for all of my SATA drives. I noticed the upgrade changed the kernel to "Linux debian 2.6.26-2-amd64" (do have 64bit CPU).Once loaded to a command prompt I can assemble my raid 6 array with the command "mdadm --assemble /dev/sda to sdd" then mount it with mount -a. But transfers to the array areorribly slow ~1mbs.Upon reboot i get the same errors and have to assemble my array every time
View 1 Replies
View Related
May 12, 2011
I have installed RHEL5U5 but I've also tested RHEL5U4 and the problem is: Once configured the multipath, I restart the server,and with the command-ll mutipath I can see the LUN presented correctly:
[Code].....
View 2 Replies
View Related
Dec 11, 2010
since ever, i face the problem, that i cannot continue using my computer properly after i suspended it to ram and wake it up again.Fedora 14, 64-bit, most recent and stable kernel
mainly KDEMotherboard: Gigabyte P55A-UD4Hard DrivesSAMSUNG HD501LJSAMSUNG SP1213NI can figure out, that everything works fine, but with a severe delay.Having a look at my led, signifying hard drive activity, it flases once a second for approx 10ms. frequency is very constant.so i performed several checks:firstly, hdparm -i, there is no difference before and after sleeping mode:
Code:
[root@bobbl-pc bobbl]# hdparm -i /dev/sda
/dev/sda:
[code]...
View 1 Replies
View Related
Jul 27, 2010
I need to set up 1 disk, the disk is presented to the server on both fibre paths from the SAN. This is the current setup seen:
[Code]..
View 1 Replies
View Related
Mar 24, 2011
I have and RHEL5.4 conecceted to an SAN HSV200. When I present a LUN to it, In system-confg-lvm gui appears a device like /dev/mapth/mapth0 (that�s the correct device to multipath) and other like /dev/sda or /dev/dm-N, as uninitialized entities. How could I configure to not display this device (/dev/sda) and prevent human errors?
This is the things I need:
1- I want (if it is posible...) to disappear the part where it says "uninitialized entities"
2- Is It necesary mantain the partitioning done to default (like /dev/mapper/mpath0p1) with device-mapper-multipath (kpartx command)? or Can I delete it without future problems?
This is the actual configuration that I have:
1- /etc/lvm/lvm.conf
Code:
preferred_names = [ "^/dev/mpath/", "^/dev/mapper/mpath", "^/dev/[hs]d" ]
filter = [ "a/dev/mapper/mpath.*/","a/dev/cciss/.*/","r/.*/" ]
2- /etc/multipath.conf
Code:
#blacklist {
# devnode "*"
#}
[code].....
View 3 Replies
View Related
Jan 10, 2010
our oracle(oracle 10G) server got a serious problem, the OS is SLES 9 SP4, two HBA adapters(QLogic 2340) with NetApp FC Storage(FAS3140), we have installed multipath-tools on this server and ensure that physical connection and swith configuration are correct, but we only can see a path
:~ # multipath -ll
360a98000572d4d414b4a522d7a62376d dm-2 NETAPP,LUN
[size=200G][features=0][hwhandler=0]
[code]....
View 3 Replies
View Related
Apr 5, 2011
I upgraded a DL580 G4 server from Red Hat 4.4 to 5.2 (2.6.18-92.el5 #1 SMP Tue Apr 29 13:16:15 EDT 2008 x86_64 x86_64 x86_64 GNU/Linux) and now when I run the multipath - l command I only see one path. There are 4 LUNs being presented on each fibre channel, so I should see sda - sdh in /dev, but I only see sda - sdd.
RPMs
device-mapper-1.02.24-1.el5
device-mapper-event-1.02.24-1.el5
device-mapper-multipath-0.4.7-17.el5
hp_qla2x00src-8.02.11-6
View 1 Replies
View Related
Apr 9, 2010
I am working on a server which connected to two EMC storage box. when i am checking the multipath information of a disk its showing as 8 paths. how to understand 8 paths for a single disk.
[Code]....
View 1 Replies
View Related
Apr 7, 2010
There is a server with two HBA QLA2460, connected via SAN to the SE9980 disk array. RHEL 5.4 (x86_64) is running on the server. In the SAN two zones are set up: HBA1-port1 9980; HBA2 - port2 SE9980. As a result, the server can see:
[Root @ c5 ~] # multipath-ll
mpath1 (1HITACHI_R45028F9022F) dm-2 HITACHI, OPEN-E
[Size = 14G] [features = 1 queue_if_no_path] [hwhandler = 0] [rw]
_ Round-robin 0 [prio = 0] [active]
_ 3:0:0:0 sdb 8:16 [active] [ready]
_ 4:0:0:0 sdc 8:32 [active] [ready]
_ 3:0:0:16384 sdd 8:48 [active] [ready]
[Code]...
Question: Why OS sees eight ways instead of two? Why do extra ways have such strange LUN (16384, 32768, etc.)?
View 1 Replies
View Related
Apr 17, 2011
after some yum update on my CentOS 5.6 server I noticed this error when the server is booting:
"dm multipath kernel driver not loaded"
View 10 Replies
View Related
Apr 23, 2010
I am familiar with active to active cluster failover and active to pasive failover, but what is network failover?
View 1 Replies
View Related
Oct 14, 2015
After mapping iscsi storage from my netapp and scanning for new devices I run multipath -v2 to create the multipath device handle under /dev/mapping/. This normally takes about 2 seconds or less on every other linux distribution I use. On Stretch it takes a little over 3 minutes. I have tried several different versions of multipath.conf but the result is always the same.
My most recent multipath.conf file is available here : [URL] ....
and the output of multipath -v4 is available here : [URL] ....
View 0 Replies
View Related
