ARP routing issue
Hello lists,
Sorry to bother you with this probably beat-to-death issue, but I can't figure out the solution to the problem.
It's perfectly described in this archive thread:
http://www.uwsg.iu.edu/hypermail/linux/net/0308.1/0071.html
Basically it comes down to this:
I have an IBM server running RH ES, kernel 2.4.9-e.49. It has two interfaces:
eth0 10.0.22.xxx
eth1 10.0.24.xxx
default gateway is set to 10.0.22.1, on eth0.
Problem is, if I try to ping from another network (10.216.0.xx) to 10.0.24.xx, i see the following ARP request:
arp who-has 10.0.22.1 tell 10.0.24.xx
which, imo, is wrong.
I know it has to do with the default gatway, but I can't devise a way to make it actually _WORK_.
Any pointers?
Thanks.
Jan
Signature · 1 line
No one can feel as helpless as the owner of a sick goldfish.
Quoted 6 lines · Jan De Luyck
Problem is, if I try to ping from another network (10.216.0.xx) to 10.0.24.xx, i see the following ARP request:
arp who-has 10.0.22.1 tell 10.0.24.xx
which, imo, is wrong.
With the info you've given it could be right or wrong. Can you provide a mini plumbing diagram to go with it. Who is arping for what too ?
Quoted 10 lines · Alan Cox
Quoted 6 lines · Jan De Luyck
Problem is, if I try to ping from another network (10.216.0.xx) to 10.0.24.xx, i see the following ARP request:
arp who-has 10.0.22.1 tell 10.0.24.xx
which, imo, is wrong.
With the info you've given it could be right or wrong. Can you provide a mini plumbing diagram to go with it. Who is arping for what too ?
Hello Alan,
Here's the plumbing schematic, in ASCII. For readability sake i've also included a png image.
--------------- 10.0.22.0 / 255.255.255.0 -----------------|
| |
eth0 | 10.0.22.x |
------------------------ |
| IBM Server | |
------------------------ |
eth1 | 10.0.24.x |
| |
--------------- 10.0.24.x / 255.255.255.0 -----------| |
| |
10.0.24.1 | | 10.0.22.1
--------------------------
| CISCO Switching router |
--------------------------
|
|
---------- 10.216.0.0 / 255.255.0.0 ----------------
|
| 10.216.x.x
------------------
| Client station |
------------------I am pinging (icmp ping) from client station to eth1 of the server.
Jan
Signature · 2 lines
Q: What do you call a boomerang that doesn't come back? A: A stick.
Hello,
Quoted 13 lines · Jan De Luyck
http://www.uwsg.iu.edu/hypermail/linux/net/0308.1/0071.html
Basically it comes down to this:
I have an IBM server running RH ES, kernel 2.4.9-e.49. It has two interfaces:
eth0 10.0.22.xxx
eth1 10.0.24.xxxdefault gateway is set to 10.0.22.1, on eth0.
Problem is, if I try to ping from another network (10.216.0.xx) to 10.0.24.xx, i see the following ARP request:
Quoted 6 lines
arp who-has 10.0.22.1 tell 10.0.24.xx
which, imo, is wrong.
I know it has to do with the default gatway, but I can't devise a way to make it actually _WORK_.
Not wrong but it is one of the possible valid requests. If it
is ignored from other boxes in your setup then you can look
at new kernels. 2.4.26 and 2.6.4 come with new sysctl flags for ARP. arp_filter filters incoming requests but you can use arp_announce to control the source IP when sending requests, eg. in IBM server you can set /proc/sys/net/ipv4/conf/eth*/arp_announce to 1 or 2 or even just /proc/sys/net/ipv4/conf/all/arp_announce to 1 or 2
Quoted 1 line
Any pointers?
See Documentation/networking/ip-sysctl.txt for more info.
Quoted 3 lines
Thanks.
Jan
Regards
--
Julian Anastasov <[email protected]>
Quoted 16 lines · Julian Anastasov
Hello,
Quoted 13 lines · Jan De Luyck
http://www.uwsg.iu.edu/hypermail/linux/net/0308.1/0071.html
Basically it comes down to this:
I have an IBM server running RH ES, kernel 2.4.9-e.49. It has two
interfaces: eth0 10.0.22.xxx
eth1 10.0.24.xxxdefault gateway is set to 10.0.22.1, on eth0.
Problem is, if I try to ping from another network (10.216.0.xx) to 10.0.24.xx, i see the following ARP request:
Quoted 8 lines
Quoted 6 lines
arp who-has 10.0.22.1 tell 10.0.24.xx
which, imo, is wrong.
I know it has to do with the default gatway, but I can't devise a way to make it actually _WORK_.
Not wrong but it is one of the possible valid requests.
Yes, I've gathered that. Yet, it seems very strange, and I want to _change_ that behaviour.
Quoted 7 lines
If it
is ignored from other boxes in your setup then you can look
at new kernels. 2.4.26 and 2.6.4 come with new sysctl flags for ARP.arp_filterfilters incoming requests but you can usearp_announceto control the source IP when sending requests, eg. in IBM server you can set/proc/sys/net/ipv4/conf/eth*/arp_announceto 1 or 2 or even just/proc/sys/net/ipv4/conf/all/arp_announceto 1 or 2
Hmm. Point is, for certification/support reasons we'd rather stick to the RedHat ES supplied kernels instead of starting off with one of our own. I can't go off running non-checked-to-be-stable (in a business POV) code on mission-critical systems. (2.6.10 is stable enough in _my_ POV, but well..)
Jan
Signature · 2 lines
Serfs up! -- Spartacus
vfs and paging error
In reply to Alan CoxAll,
After few days(5-6) of running my IBM-NETVISTA system with Kernel-2.4.26, I
get following errors
[sumit@linux139 sumit]$ cut -d" " -f5- /tmp/messages
kernel: <1>Unable to handle kernel paging request at virtual address 57e34010
kernel: printing eip:
kernel: c014b236
kernel: *pde = 00000000
kernel: Oops: 0000
kernel: CPU: 0
kernel: EIP: 0010:[expand_fdset+198/480] Not tainted
kernel: EFLAGS: 00010207
kernel: eax: 57e34010 ebx: 57e34000 ecx: 0000ffff edx: 01cb142c
kernel: esi: cf898328 edi: c8b9bf84 ebp: c8b9be8c esp: c8b9be64
kernel: ds: 0018 es: 0018 ss: 0018
kernel: Process ifconfig (pid: 15052, stackpage=c8b9b000)
kernel: Stack: 00000000 c0125ed1 57e34010 dfe34010 c7e0500a 01cb142c 00000004 c8b9bf04
kernel: cf898328 c8b9bf84 c8b9beac c0141f34 cf898328 c8b9bf04cfd342f8 ffffffec
kernel: dca0b66c c8b9bf04 c8b9bf2c c0142838 cf898328 c8b9bf04
00000000 00000246
kernel: Call Trace: [get_unmapped_area+145/288] [vfs_rename_dir+180/1312]
[sys_rename+184/480] [__change_page_attr+172/368] [vfs_link+48/272]
kernel: [page_follow_link+113/397] [.text.lock.namei+336/531]
[fput+76/240] [invalidate_bdev+11/320] [device_not_available_emulate+15/16] kernel:
kernel: Code: 8b 00 89 45 e0 39 53 44 75 78 8b 45 08 39 43 0c 75 70 8b 40
kernel: <1>Unable to handle kernel paging request at virtual address 57e34010
kernel: printing eip:
kernel: c014b236
kernel: *pde = 00000000
kernel: Oops: 0000
kernel: CPU: 0
kernel: EIP: 0010:[expand_fdset+198/480] Not tainted
kernel: EFLAGS: 00010207
kernel: eax: 57e34010 ebx: 57e34000 ecx: 0000ffff edx: 01cb142c
kernel: esi: cf898328 edi: c8b9bf84 ebp: c8b9be8c esp: c8b9be64
kernel: ds: 0018 es: 0018 ss: 0018
kernel: Process ifconfig (pid: 15054, stackpage=c8b9b000)
kernel: Stack: 00000000 c0125ed1 57e34010 dfe34010 d957c00a 01cb142c 00000004 c8b9bf04
kernel: cf898328 c8b9bf84 c8b9beac c0141f34 cf898328 c8b9bf04cda262f8 ffffffec
ifup: SIOCADDRT: Network is unreachable
kernel: dca0b66c c8b9bf04 c8b9bf2c c0142838 cf898328 c8b9bf04
00000000 00000246
kernel: Call Trace: [get_unmapped_area+145/288] [vfs_rename_dir+180/1312]
[sys_rename+184/480] [__change_page_attr+172/368] [vfs_link+48/272]
kernel: [page_follow_link+113/397] [.text.lock.namei+336/531]
[fput+76/240] [device_not_available_emulate+15/16]
kernel:
kernel: Code: 8b 00 89 45 e0 39 53 44 75 78 8b 45 08 39 43 0c 75 70 8b 40
kernel: <1>Unable to handle kernel paging request at virtual address 57e34010
kernel: printing eip:
kernel: c014b236
kernel: *pde = 00000000
kernel: Oops: 0000
kernel: CPU: 0
kernel: EIP: 0010:[expand_fdset+198/480] Not tainted
kernel: EFLAGS: 00010207
kernel: eax: 57e34010 ebx: 57e34000 ecx: 0000ffff edx: 01cb142c
kernel: esi: cf898328 edi: c8b3bf84 ebp: c8b3be8c esp: c8b3be64
kernel: ds: 0018 es: 0018 ss: 0018
kernel: Process ifconfig (pid: 15082, stackpage=c8b3b000)
kernel: Stack: 00000000 c0125ed1 57e34010 dfe34010 c404800a 01cb142c 00000004 c8b3bf04
kernel: cf898328 c8b3bf84 c8b3beac c0141f34 cf898328 c8b3bf04cfd342f8 ffffffec
kernel: dca0b66c c8b3bf04 c8b3bf2c c0142838 cf898328 c8b3bf04
00000000 00000246
kernel: Call Trace: [get_unmapped_area+145/288] [vfs_rename_dir+180/1312]
[sys_rename+184/480] [__change_page_attr+172/368] [vfs_link+48/272]
kernel: [page_follow_link+113/397] [.text.lock.namei+336/531]
[fput+76/240] [probe_irq_off+130/144] [device_not_available_emulate+15/16] kernel:
kernel: Code: 8b 00 89 45 e0 39 53 44 75 78 8b 45 08 39 43 0c 75 70 8b 40
network: Bringing up interface lo succeeded
kernel: <1>Unable to handle kernel paging request at virtual address 57e34010
kernel: printing eip:
kernel: c014b236
kernel: *pde = 00000000
kernel: Oops: 0000
kernel: CPU: 0
kernel: EIP: 0010:[expand_fdset+198/480] Not tainted
kernel: EFLAGS: 00010207
kernel: eax: 57e34010 ebx: 57e34000 ecx: 0000ffff edx: 01cb142c
kernel: esi: cf898328 edi: c8b9bf84 ebp: c8b9be8c esp: c8b9be64
kernel: ds: 0018 es: 0018 ss: 0018
kernel: Process ifconfig (pid: 15114, stackpage=c8b9b000)
kernel: Stack: 00000000 c0125ed1 57e34010 dfe34010 d576b00a 01cb142c 00000004 c8b9bf04
kernel: cf898328 c8b9bf84 c8b9beac c0141f34 cf898328 c8b9bf04cda262f8 ffffffec
kernel: dca0b66c c8b9bf04 c8b9bf2c c0142838 cf898328 c8b9bf04
00000000 00000246
kernel: Call Trace: [get_unmapped_area+145/288] [vfs_rename_dir+180/1312]
[sys_rename+184/480] [__change_page_attr+172/368] [vfs_link+48/272]
kernel: [page_follow_link+113/397] [.text.lock.namei+336/531]
[fput+76/240] [invalidate_bdev+11/320] [device_not_available_emulate+15/16] kernel:
kernel: Code: 8b 00 89 45 e0 39 53 44 75 78 8b 45 08 39 43 0c 75 70 8b 40
kernel: <1>Unable to handle kernel paging request at virtual address 57e34010
kernel: printing eip:
kernel: c014b236
kernel: *pde = 00000000
kernel: Oops: 0000
kernel: CPU: 0
kernel: EIP: 0010:[expand_fdset+198/480] Not tainted
kernel: EFLAGS: 00010207
kernel: eax: 57e34010 ebx: 57e34000 ecx: 0000ffff edx: 01cb142c
kernel: esi: cf898328 edi: c8b9bf84 ebp: c8b9be8c esp: c8b9be64
kernel: ds: 0018 es: 0018 ss: 0018
kernel: Process ifconfig (pid: 15120, stackpage=c8b9b000)
kernel: Stack: 00000000 0000000e 57e34010 dfe34010 c83bc00a 01cb142c 00000004 c8b9bf04
kernel: cf898328 c8b9bf84 c8b9beac c0141f34 cf898328 c8b9bf04c76032f8 ffffffec
kernel: dca0b66c c8b9bf04 c8b9bf2c c0142838 cf898328 c8b9bf04
00000000 00000246
kernel: Call Trace: [vfs_rename_dir+180/1312] [sys_rename+184/480]
[__change_page_attr+172/368] [vfs_link+48/272] [page_follow_link+113/397]
kernel: [.text.lock.namei+336/531] [fput+76/240]
[device_not_available_emulate+15/16]
kernel:
kernel: Code: 8b 00 89 45 e0 39 53 44 75 78 8b 45 08 39 43 0c 75 70 8b 40
ifup: SIOCADDRT: Network is down
ifup: SIOCADDRT: Network is unreachable
kernel: <1>Unable to handle kernel paging request at virtual address 57e34010
kernel: printing eip:
kernel: c014b236
kernel: *pde = 00000000
kernel: Oops: 0000
kernel: CPU: 0
kernel: EIP: 0010:[expand_fdset+198/480] Not tainted
kernel: EFLAGS: 00010207
kernel: eax: 57e34010 ebx: 57e34000 ecx: 0000ffff edx: 01cb142c
kernel: esi: cf898328 edi: c84d1f84 ebp: c84d1e8c esp: c84d1e64
kernel: ds: 0018 es: 0018 ss: 0018
kernel: Process ifconfig (pid: 15148, stackpage=c84d1000)
kernel: Stack: 00000000 c0125ed1 57e34010 dfe34010 d926200a 01cb142c 00000004 c84d1f04
kernel: cf898328 c84d1f84 c84d1eac c0141f34 cf898328 c84d1f04d000e2f8 ffffffec
kernel: dca0b66c c84d1f04 c84d1f2c c0142838 cf898328 c84d1f04
00000000 00000246
kernel: Call Trace: [get_unmapped_area+145/288] [vfs_rename_dir+180/1312]
[sys_rename+184/480] [__change_page_attr+172/368] [vfs_link+48/272]
kernel: [page_follow_link+113/397] [.text.lock.namei+336/531]
[fput+76/240] [invalidate_bdev+11/320] [device_not_available_emulate+15/16] kernel:
kernel: Code: 8b 00 89 45 e0 39 53 44 75 78 8b 45 08 39 43 0c 75 70 8b 40
[sumit@linux139 sumit]$ cat /proc/cmdline BOOT_IMAGE=linux-up ro root=305 BOOT_FILE=/boot/vmlinuz-2.4.26-3 console=ttyS0 acpi=off pci=noacpi acpi=noirq noapic
I've to run this system without ACPI and APIC because suddenly after some
2-3 days of uptime my system clock run in very unsyncronized way; sometimes even upto 3 times faster. I reported this timer issue before but couldn't manage to implement the suggession of verifying this with 2.4.27 or even 2.4.28.
Thus finally workaround of disabling all acpi and apic issues worked
against timer issue but then this another problem stucks.
Thanks for your time,
-- Sumit