1828668
Members
2032
Online
109984
Solutions
Hi,
The problem is that for one reason or another you have exceeded the total number of ext_bus entries you can have on a system (at 11iv1 or 11iv2 - this isn't a limit with 11iv3). This is either because:
a) You have A LOT of LUNs on this system -or-
b) the config of the LUNs has changed several times such that there are now many 'ghost' entries in the /etc/ioconfig file
You can tell which by running 'ioscan -FC ext_bus | wc -l' - if you have more than 255 then you really do need to reorganise the way you present storage to this system to cut down the number of buses they're presented on (what sort of storage do you have?). If its significantly less than 255, you just have a lot of old 'ghost' entries in your ioconfig file.
Either way the ioconfig file now contains over 255 entries for ext_bus, and can't add any more - so you need to get that resolved.
You're gonna have to get a call open with HP support to resolve this, as you will need a utility called ioconfig_dump which can only be provided by a RC engineer. Ask the engineer to reference doc ID emr_na-c01007568-2. Be prepared for some reboots.
If you *don't* have support and you really can't see 255 ext_bus entries in your ioscan, then the only course of action you have available is to go through the risky process of re-creating the ioconfig file from scratch, which would basically involve:
i) shutting down your apps and unmounting all but your root VG.
ii) Creating map files for all your VGs using 'vgexport -p -s -m vgXX.map vgXX'
iii) saving a copy of the output of 'ioscan -fn' so you have a picture of how instance numbers matched to hardware paths before the change.
iv) saving copies of the output of 'lanscan' and 'netstat -in' so you know how LAN instance numbers match up to IP addresses.
v) Moving out the old ioconfig file /etc/ioconfig to a safe location on the root filesystem
vi) running 'ioinit -c' to create a new ioconfig
v) Rebooting 'reboot -rs'
vi) If the disks in the root volume group have had their instance ID changed (a definite possibility), then the boot will probably fail with an LVM problem, which will mean rebootong to LVM maintenance mode and re-creating the root volume group using vgexport/vgimport/lvlnboot.
vii) Once the root volume group is sorted you'll probably be able to get to multi-user mode, then you'll have to sort the other volume groups using vgexport/vgimport.
viii) You'll then need to look and see if your LAN instance numbers have changed - if they have you'll need to adjust /etc/rc.config.d/netconf appropriately to make sure IPs are bourebootingnd to the right LAN cards.
ix) and of course there could be other devices that have changed as a result of this - compare against what the new and old ioscan looked like.
If that doesn't all make immediate sense to you and you know what I'm talking about, I would seriously advise against going down this option.
HTH
Duncan
I am an HPE Employee
The problem is that for one reason or another you have exceeded the total number of ext_bus entries you can have on a system (at 11iv1 or 11iv2 - this isn't a limit with 11iv3). This is either because:
a) You have A LOT of LUNs on this system -or-
b) the config of the LUNs has changed several times such that there are now many 'ghost' entries in the /etc/ioconfig file
You can tell which by running 'ioscan -FC ext_bus | wc -l' - if you have more than 255 then you really do need to reorganise the way you present storage to this system to cut down the number of buses they're presented on (what sort of storage do you have?). If its significantly less than 255, you just have a lot of old 'ghost' entries in your ioconfig file.
Either way the ioconfig file now contains over 255 entries for ext_bus, and can't add any more - so you need to get that resolved.
You're gonna have to get a call open with HP support to resolve this, as you will need a utility called ioconfig_dump which can only be provided by a RC engineer. Ask the engineer to reference doc ID emr_na-c01007568-2. Be prepared for some reboots.
If you *don't* have support and you really can't see 255 ext_bus entries in your ioscan, then the only course of action you have available is to go through the risky process of re-creating the ioconfig file from scratch, which would basically involve:
i) shutting down your apps and unmounting all but your root VG.
ii) Creating map files for all your VGs using 'vgexport -p -s -m vgXX.map vgXX'
iii) saving a copy of the output of 'ioscan -fn' so you have a picture of how instance numbers matched to hardware paths before the change.
iv) saving copies of the output of 'lanscan' and 'netstat -in' so you know how LAN instance numbers match up to IP addresses.
v) Moving out the old ioconfig file /etc/ioconfig to a safe location on the root filesystem
vi) running 'ioinit -c' to create a new ioconfig
v) Rebooting 'reboot -rs'
vi) If the disks in the root volume group have had their instance ID changed (a definite possibility), then the boot will probably fail with an LVM problem, which will mean rebootong to LVM maintenance mode and re-creating the root volume group using vgexport/vgimport/lvlnboot.
vii) Once the root volume group is sorted you'll probably be able to get to multi-user mode, then you'll have to sort the other volume groups using vgexport/vgimport.
viii) You'll then need to look and see if your LAN instance numbers have changed - if they have you'll need to adjust /etc/rc.config.d/netconf appropriately to make sure IPs are bourebootingnd to the right LAN cards.
ix) and of course there could be other devices that have changed as a result of this - compare against what the new and old ioscan looked like.
If that doesn't all make immediate sense to you and you know what I'm talking about, I would seriously advise against going down this option.
HTH
Duncan
I am an HPE Employee
yikes, what did the spell checker do there?
Item viii) should read:
viii) You'll then need to look and see if your LAN instance numbers have changed - if they have you'll need to adjust /etc/rc.config.d/netconf appropriately to make sure IPs are bound to the right LAN cards.
Duncan
I am an HPE Employee
Item viii) should read:
viii) You'll then need to look and see if your LAN instance numbers have changed - if they have you'll need to adjust /etc/rc.config.d/netconf appropriately to make sure IPs are bound to the right LAN cards.
Duncan
I am an HPE Employee