Proxmox VE 4 : iscsi and multipath

phb19

New Member
Oct 9, 2015
2
0
1
Hello,


I use proxmox ve since two years and it's a great product. Actually i use the 3.3 version.

My configuration : 8 servers and 3 Equallogic ( ISCSI array stoage ). All my Vms are KVM.
I use two network card for iscsi network and multipath. All is Okay

I actually try to install proxmox ve 4. I use a new server. I have a problem with multipath daemon/udev.

When i connect a lun, multipath failed. It looks like multipath run before iscsi is completly ready

This is what it happen when i connect a lun :

In syslog i have :
Code:
Oct 13 10:40:37 gyndine-l10 kernel: [74396.332084] scsi host48: iSCSI Initiator over TCP/IP
Oct 13 10:40:37 gyndine-l10 kernel: [74396.342094] scsi host49: iSCSI Initiator over TCP/IP
Oct 13 10:40:37 gyndine-l10 kernel: [74396.344810]  connection41:0: detected conn error (1020)
Oct 13 10:40:37 gyndine-l10 kernel: [74396.352928]  connection42:0: detected conn error (1020)
Oct 13 10:40:39 gyndine-l10 iscsid: Login authentication failed with target iqn.2001-05.com.equallogic:0-8a0906-ceaa82a0a-f56005cf25f5512a-testiscsi
Oct 13 10:40:39 gyndine-l10 iscsid: Login authentication failed with target iqn.2001-05.com.equallogic:0-8a0906-ceaa82a0a-f56005cf25f5512a-testiscsi
Oct 13 10:40:39 gyndine-l10 kernel: [74397.903492] scsi 48:0:0:0: Direct-Access     EQLOGIC  100E-00          7.0  PQ: 0 ANSI: 5
Oct 13 10:40:39 gyndine-l10 kernel: [74397.903970] scsi 49:0:0:0: Direct-Access     EQLOGIC  100E-00          7.0  PQ: 0 ANSI: 5
Oct 13 10:40:39 gyndine-l10 kernel: [74397.904213] sd 48:0:0:0: Attached scsi generic sg3 type 0
Oct 13 10:40:39 gyndine-l10 kernel: [74397.904529] sd 48:0:0:0: [sdb] 109056000 512-byte logical blocks: (55.8 GB/52.0 GiB)
Oct 13 10:40:39 gyndine-l10 kernel: [74397.905076] sd 49:0:0:0: [sdc] 109056000 512-byte logical blocks: (55.8 GB/52.0 GiB)
Oct 13 10:40:39 gyndine-l10 kernel: [74397.905087] sd 49:0:0:0: Attached scsi generic sg4 type 0
Oct 13 10:40:39 gyndine-l10 kernel: [74397.905231] sd 48:0:0:0: [sdb] Write Protect is off
Oct 13 10:40:39 gyndine-l10 kernel: [74397.905237] sd 48:0:0:0: [sdb] Mode Sense: 81 00 00 00
Oct 13 10:40:39 gyndine-l10 kernel: [74397.905445] sd 48:0:0:0: [sdb] Write cache: disabled, read cache: enabled, doesn't support DPO or FUA
Oct 13 10:40:39 gyndine-l10 kernel: [74397.905838] sd 49:0:0:0: [sdc] Write Protect is off
Oct 13 10:40:39 gyndine-l10 kernel: [74397.905845] sd 49:0:0:0: [sdc] Mode Sense: 81 00 00 00
Oct 13 10:40:39 gyndine-l10 kernel: [74397.906069] sd 49:0:0:0: [sdc] Write cache: disabled, read cache: enabled, doesn't support DPO or FUA
Oct 13 10:40:39 gyndine-l10 kernel: [74397.911308]  sdb: sdb1 sdb2 sdb3
Oct 13 10:40:39 gyndine-l10 kernel: [74397.913243] sd 48:0:0:0: [sdb] Attached SCSI disk
Oct 13 10:40:39 gyndine-l10 kernel: [74397.915077]  sdc: sdc1 sdc2 sdc3
Oct 13 10:40:39 gyndine-l10 kernel: [74397.916958] sd 49:0:0:0: [sdc] Attached SCSI disk
Oct 13 10:40:40 gyndine-l10 iscsid: Connection41:0 to [target: iqn.2001-05.com.equallogic:0-8a0906-ceaa82a0a-f56005cf25f5512a-testiscsi, portal: 172.23.230.254,3260] through [iface: iscsi0] is operational now
Oct 13 10:40:40 gyndine-l10 iscsid: Connection42:0 to [target: iqn.2001-05.com.equallogic:0-8a0906-ceaa82a0a-f56005cf25f5512a-testiscsi, portal: 172.23.230.254,3260] through [iface: iscsi1] is operational now

During 30 seconds i have two process multipath at 15-20 % :

Code:
PID USER      PR  NI    VIRT    RES    SHR S  %CPU %MEM     TIME+ COMMAND

 11033 root      20   0   30712   3316   2524 D  17.8  0.0   0:00.34 multipath                                                                                                                                                                  

 11034 root      20   0   30712   3220   2448 D  17.8  0.0   0:00.34 multipath

And after 30 seconds :
Code:
 Oct 13 10:41:10 gyndine-l10 systemd-udevd[11014]: timeout '/sbin/multipath -v0 /dev/sdb'
 Oct 13 10:41:10 gyndine-l10 systemd-udevd[11015]: timeout '/sbin/multipath -v0 /dev/sdc'
 Oct 13 10:41:11 gyndine-l10 systemd-udevd[11014]: timeout: killing '/sbin/multipath -v0 /dev/sdb' [11033]
 Oct 13 10:41:11 gyndine-l10 systemd-udevd[11014]: '/sbin/multipath -v0 /dev/sdb' [11033] terminated by signal 9 (Killed)
 Oct 13 10:41:11 gyndine-l10 systemd-udevd[11014]: timeout '/lib/udev/hdparm'
 Oct 13 10:41:11 gyndine-l10 systemd-udevd[11015]: timeout: killing '/sbin/multipath -v0 /dev/sdc' [11034]
 Oct 13 10:41:11 gyndine-l10 multipathd: sdb: add path (uevent)
 Oct 13 10:41:11 gyndine-l10 multipathd: mpatha: failed in domap for addition of new path sdb
 Oct 13 10:41:11 gyndine-l10 systemd-udevd[11015]: '/sbin/multipath -v0 /dev/sdc' [11034] terminated by signal 9 (Killed)
 Oct 13 10:41:11 gyndine-l10 systemd-udevd[11015]: timeout '/lib/udev/hdparm'
 Oct 13 10:41:11 gyndine-l10 multipathd: uevent trigger error
 Oct 13 10:41:11 gyndine-l10 multipathd: sdc: add path (uevent)
 Oct 13 10:41:11 gyndine-l10 multipathd: mpatha: failed in domap for addition of new path sdc
 Oct 13 10:41:11 gyndine-l10 multipathd: uevent trigger error

If i disable the /lib/udev/rules.d/60-multipath.rules, i have the same problem ( but without the two process multipath at 15-20%) :
Code:
Oct 13 10:46:03 gyndine-l10 kernel: [74721.444886] scsi host50: iSCSI Initiator over TCP/IP
 Oct 13 10:46:03 gyndine-l10 kernel: [74721.455195] scsi host51: iSCSI Initiator over TCP/IP
 Oct 13 10:46:03 gyndine-l10 kernel: [74721.457892]  connection43:0: detected conn error (1020)
 Oct 13 10:46:03 gyndine-l10 kernel: [74721.466012]  connection44:0: detected conn error (1020)
 Oct 13 10:46:04 gyndine-l10 iscsid: Login authentication failed with target iqn.2001-05.com.equallogic:0-8a0906-ceaa82a0a-f56005cf25f5512a-testiscsi
 Oct 13 10:46:04 gyndine-l10 iscsid: Login authentication failed with target iqn.2001-05.com.equallogic:0-8a0906-ceaa82a0a-f56005cf25f5512a-testiscsi
 Oct 13 10:46:04 gyndine-l10 kernel: [74722.998303] scsi 50:0:0:0: Direct-Access     EQLOGIC  100E-00          7.0  PQ: 0 ANSI: 5
 Oct 13 10:46:04 gyndine-l10 kernel: [74722.998776] scsi 51:0:0:0: Direct-Access     EQLOGIC  100E-00          7.0  PQ: 0 ANSI: 5
 Oct 13 10:46:04 gyndine-l10 kernel: [74722.999150] sd 50:0:0:0: [sdb] 109056000 512-byte logical blocks: (55.8 GB/52.0 GiB)
 Oct 13 10:46:04 gyndine-l10 kernel: [74722.999209] sd 50:0:0:0: Attached scsi generic sg3 type 0
 Oct 13 10:46:04 gyndine-l10 kernel: [74722.999475] sd 51:0:0:0: Attached scsi generic sg4 type 0
 Oct 13 10:46:04 gyndine-l10 kernel: [74722.999666] sd 51:0:0:0: [sdc] 109056000 512-byte logical blocks: (55.8 GB/52.0 GiB)
 Oct 13 10:46:04 gyndine-l10 kernel: [74722.999964] sd 50:0:0:0: [sdb] Write Protect is off
 Oct 13 10:46:04 gyndine-l10 kernel: [74722.999992] sd 50:0:0:0: [sdb] Mode Sense: 81 00 00 00
 Oct 13 10:46:04 gyndine-l10 kernel: [74723.000326] sd 50:0:0:0: [sdb] Write cache: disabled, read cache: enabled, doesn't support DPO or FUA
 Oct 13 10:46:04 gyndine-l10 kernel: [74723.000621] sd 51:0:0:0: [sdc] Write Protect is off
 Oct 13 10:46:04 gyndine-l10 kernel: [74723.000628] sd 51:0:0:0: [sdc] Mode Sense: 81 00 00 00
 Oct 13 10:46:04 gyndine-l10 kernel: [74723.000888] sd 51:0:0:0: [sdc] Write cache: disabled, read cache: enabled, doesn't support DPO or FUA
 Oct 13 10:46:04 gyndine-l10 kernel: [74723.009825]  sdb: sdb1 sdb2 sdb3
 Oct 13 10:46:04 gyndine-l10 kernel: [74723.011669] sd 50:0:0:0: [sdb] Attached SCSI disk
 Oct 13 10:46:04 gyndine-l10 kernel: [74723.013591]  sdc: sdc1 sdc2 sdc3
 Oct 13 10:46:04 gyndine-l10 kernel: [74723.015377] sd 51:0:0:0: [sdc] Attached SCSI disk
 Oct 13 10:46:05 gyndine-l10 multipathd: sdb: add path (uevent)
 Oct 13 10:46:05 gyndine-l10 multipathd: mpatha: failed in domap for addition of new path sdb
 Oct 13 10:46:05 gyndine-l10 multipathd: uevent trigger error
 Oct 13 10:46:05 gyndine-l10 multipathd: sdc: add path (uevent)
 Oct 13 10:46:05 gyndine-l10 multipathd: mpatha: failed in domap for addition of new path sdc
 Oct 13 10:46:05 gyndine-l10 iscsid: Connection43:0 to [target: iqn.2001-05.com.equallogic:0-8a0906-ceaa82a0a-f56005cf25f5512a-testiscsi, portal: 172.23.230.254,3260] through [iface: iscsi0] is operational now
 Oct 13 10:46:05 gyndine-l10 iscsid: Connection44:0 to [target: iqn.2001-05.com.equallogic:0-8a0906-ceaa82a0a-f56005cf25f5512a-testiscsi, portal: 172.23.230.254,3260] through [iface: iscsi1] is operational now
 Oct 13 10:46:05 gyndine-l10 multipathd: uevent trigger error
How can i solve this problem ?


My configuration of multipath seems to be good because if un run manually multipath after iscsi login ok, it's ok :
Code:
    multipath

 create: mpatha (36090a0a8a082aace2a51f525cf0560f5) undef EQLOGIC,100E-00
  size=52G features='1 queue_if_no_path' hwhandler='0' wp=undef
 `-+- policy='round-robin 0' prio=1 status=undef
   |- 50:0:0:0 sdb 8:16 undef ready running
   `- 51:0:0:0 sdc 8:32 undef ready running

and in syslog :


Code:
Oct 13 10:55:59 gyndine-l10 multipathd: mpatha: adding map

 Oct 13 10:55:59 gyndine-l10 multipathd: mpatha: event checker started
 Oct 13 10:55:59 gyndine-l10 multipathd: mpatha: devmap dm-3 registered

Thanks for your help
 
I installed sid package of multipath-tools and that works.

So it looks like a bug in multipath-tools