Debian
在引導期間啟動多路徑上的 LVM 分區的正確方法
我有成功配置 iSCSI 和多路徑的 Debian 9：
# multipath -ll /dev/mapper/mpathb mpathb (222c60001556480c6) dm-2 Promise,Vess R2600xi size=10T features='1 retain_attached_hw_handler' hwhandler='0' wp=rw |-+- policy='service-time 0' prio=1 status=active | `- 12:0:0:0 sdc 8:32 active ready running `-+- policy='service-time 0' prio=1 status=enabled `- 13:0:0:0 sdd 8:48 active ready running
/dev/mapper/mpathb是 LVM 組的一部分vg-one-100：# pvs PV VG Fmt Attr PSize PFree /dev/dm-2 vg-one-100 lvm2 a-- 10,00t 3,77t # vgs VG #PV #LV #SN Attr VSize VFree vg-one-100 1 17 0 wz--n- 10,00t 3,77t
vg-one-100組包含幾卷：# lvs LV VG Attr LSize Pool Origin Data% Meta% Move Log Cpy%Sync Convert lv-one-0-1 vg-one-100 -wi-a----- 20,00g lv-one-1-0 vg-one-100 -wi-a----- 2,41g lv-one-10-0 vg-one-100 -wi------- 20,00g lv-one-11-0 vg-one-100 -wi------- 30,00g lv-one-12-0 vg-one-100 -wi------- 2,41g lv-one-13-0 vg-one-100 -wi------- 2,41g lv-one-14-0 vg-one-100 -wi------- 2,41g lv-one-15-0 vg-one-100 -wi------- 2,41g lv-one-16-0 vg-one-100 -wi------- 2,41g lv-one-17-0 vg-one-100 -wi------- 30,00g lv-one-18-0 vg-one-100 -wi------- 30,00g lv-one-23-0 vg-one-100 -wi------- 20,00g lv-one-31-0 vg-one-100 -wi------- 20,00g lv-one-8-0 vg-one-100 -wi------- 30,00g lv-one-9-0 vg-one-100 -wi------- 20,00g lvm_images vg-one-100 -wi-a----- 5,00t lvm_system vg-one-100 -wi-a----- 1,00t我的
lvm.conf包括下一個過濾器：# grep filter /etc/lvm/lvm.conf | grep -vE '^.*#' filter = ["a|/dev/dm-*|", "r|.*|"] global_filter = ["a|/dev/dm-*|", "r|.*|"]
lvmetad被禁用：# grep use_lvmetad /etc/lvm/lvm.conf | grep -vE '^.*#' use_lvmetad = 0如果
lvmetad禁用，lvm2-activation-generator則將使用。就我而言
lvm2-activation-generator，生成了所有需要的單元文件並在引導期間執行它：# ls -1 /var/run/systemd/generator/lvm2-activation* /var/run/systemd/generator/lvm2-activation-early.service /var/run/systemd/generator/lvm2-activation-net.service /var/run/systemd/generator/lvm2-activation.service # systemctl status lvm2-activation-early.service ● lvm2-activation-early.service - Activation of LVM2 logical volumes Loaded: loaded (/etc/lvm/lvm.conf; generated; vendor preset: enabled) Active: inactive (dead) since Thu 2019-03-28 17:20:48 MSK; 3 weeks 4 days ago Docs: man:lvm2-activation-generator(8) Main PID: 897 (code=exited, status=0/SUCCESS) systemd[1]: Starting Activation of LVM2 logical volumes... systemd[1]: Started Activation of LVM2 logical volumes. root@virt1:~# systemctl status lvm2-activation-net.service ● lvm2-activation-net.service - Activation of LVM2 logical volumes Loaded: loaded (/etc/lvm/lvm.conf; generated; vendor preset: enabled) Active: inactive (dead) since Thu 2019-03-28 17:21:24 MSK; 3 weeks 4 days ago Docs: man:lvm2-activation-generator(8) Main PID: 1537 (code=exited, status=0/SUCCESS) systemd[1]: Starting Activation of LVM2 logical volumes... lvm[1537]: 4 logical volume(s) in volume group "vg-one-100" now active systemd[1]: Started Activation of LVM2 logical volumes. root@virt1:~# systemctl status lvm2-activation.service ● lvm2-activation.service - Activation of LVM2 logical volumes Loaded: loaded (/etc/lvm/lvm.conf; generated; vendor preset: enabled) Active: inactive (dead) since Thu 2019-03-28 17:20:48 MSK; 3 weeks 4 days ago Docs: man:lvm2-activation-generator(8) Main PID: 900 (code=exited, status=0/SUCCESS) systemd[1]: Starting Activation of LVM2 logical volumes... systemd[1]: Started Activation of LVM2 logical volumes.問題在於：我無法在引導期間自動啟動所有 LVM 卷，因為在通過 iSCSI 而不是多路徑設備（片段）
lvm2-activator-net.service連接（登錄）卷之後啟動卷：journalctl. . . kernel: sd 11:0:0:0: [sdc] 21474836480 512-byte logical blocks: (11.0 TB/10.0 TiB) kernel: sd 10:0:0:0: [sdb] Write cache: enabled, read cache: enabled, supports DPO and FUA kernel: sd 11:0:0:0: [sdc] Write Protect is off kernel: sd 11:0:0:0: [sdc] Mode Sense: 97 00 10 08 kernel: sd 11:0:0:0: [sdc] Write cache: enabled, read cache: enabled, supports DPO and FUA kernel: sd 10:0:0:0: [sdb] Attached SCSI disk kernel: sd 11:0:0:0: [sdc] Attached SCSI disk iscsiadm[1765]: Logging in to [iface: default, target: iqn.2012-07.com.promise:alias.tgt0000.2000000155588d75, portal: 172.16.0.151,3260] (multiple) iscsiadm[1765]: Logging in to [iface: default, target: iqn.2012-07.com.promise:alias.tgt0000.2000000155588d75, portal: 172.16.1.151,3260] (multiple) iscsiadm[1765]: Login to [iface: default, target: iqn.2012-07.com.promise:alias.tgt0000.2000000155588d75, portal: 172.16.0.151,3260] successful. iscsiadm[1765]: Login to [iface: default, target: iqn.2012-07.com.promise:alias.tgt0000.2000000155588d75, portal: 172.16.1.151,3260] successful. systemd[1]: Started Login to default iSCSI targets. systemd[1]: Starting Activation of LVM2 logical volumes... systemd[1]: Starting Activation of LVM2 logical volumes... multipathd[884]: sdb: add path (uevent) systemd[1]: Started Activation of LVM2 logical volumes. systemd[1]: Started Activation of LVM2 logical volumes. systemd[1]: Reached target Remote File Systems (Pre). systemd[1]: Mounting /var/lib/one/datastores/101... systemd[1]: Mounting /var/lib/one/datastores/100... multipathd[884]: mpathb: load table [0 21474836480 multipath 1 retain_attached_hw_handler 0 1 1 service-time 0 1 1 8:16 1] multipathd[884]: mpathb: event checker started multipathd[884]: sdb [8:16]: path added to devmap mpathb multipathd[884]: sdc: add path (uevent) multipathd[884]: mpathb: load table [0 21474836480 multipath 1 retain_attached_hw_handler 0 2 1 service-time 0 1 1 8:16 1 service-time 0 1 1 8:32 1] . . .啟動條件
lvm2-activation-net.service正確：# grep After /var/run/systemd/generator/lvm2-activation-net.service After=lvm2-activation.service iscsi.service fcoe.service如何
all在引導期間正確啟動邏輯卷？
由於您似乎只有一個物理卷，我真的很想知道在您的情況下如何發生部分啟動。它應該是全部或全部。但無論如何，這裡有幾個問題需要處理：
- 您需要持久的多路徑設備名稱。我不確定從哪裡來，但為了清楚起見
mpathb，我建議不要啟用。手動配置別名或使用儲存提供的 WWID。user_friendly_names``/etc/multipath.conf- LVM 過濾器是正則表達式，而不是 shell glob，因此您需要將語法更改為類似
filter = ["a|^/dev/mapper/222c60001556480c6$|", "r|.|"]（
global_filter對於正確的功能是可選的，但它可能會影響啟動時間。）
- 您必須延遲啟動，直到所有物理卷的多路徑設備出現。一種可能性是添加
Requires = dev-mapper-222c60001556480c6.device After = dev-mapper-222c60001556480c6.device到
/etc/systemd/system/lvm2-activation-net.service.d/wait_for_storage.conf. 另一個是創建專門的啟動服務。
- iSCSI 儲存設備（及其多路徑設備）可能需要很長時間才能出現。您可能需要創建
/etc/systemd/system/dev-mapper-222c60001556480c6.device包含[Unit] JobTimeoutSec=3min以確保 systemd 不會因為等待它而超時。如果您有多個此類設備，請使用指向通用文件的符號連結。
即使上述方法不能立即解決您的問題，它也會使調試更容易處理。祝你好運！