Re: Proxy hangs completly after an uptime of 3Weeks

You definitely have a kernel bug, or memory fault. Send your info to the
linux-kernel list as a bug report. Don't forget to include the _version_
of linux.

Sascha Schnitzler wrote:
>
> Hello again :-)
>
> Last night, my complete Proxy-Box hung up after an uptime of about 22
> Days. This is the second time this happened - the first hungup came also
> after an uptime of about 3 weeks, although only Squid wasn't running
> anymore (the rest was still alive).
>
> The box is a P200MMX with 32MB RAM and an 2,1 GB FUJITSU E-IDE Harddisk
> (not my choice :-)). It's only purpose is beeing a Web-Proxy so there are
> no other Deamons started beside Squid 1.18 (_not_ NOVM).
>
> Squid uses 1GB for it's cache and cache_mem is 8MB. According to
> cachemgr.cgi I'm using around 24MB (incl. cache_mem) of RAM for Squid. The
> highest SWAP-usage I've seen in regular use was around 3 MB. (A
> RAM-Upgrade to 64 MB is planned)
>
> The proxy isn't very high loaded - Less than 900 Connections per hour.
>
> To be complete, here's what my Kernel reported to the Syslog (I have no
> plan on the following so I'm appreciating any 'translations')
>
> I hope someone can help me and thank you very much in advance.
>
> May 19 00:51:36 proxy kernel: Warning: kfree_skb passed an skb still on a
> list (from 0088acf0).
> May 19 00:51:36 proxy kernel: Warning: kfree_skb passed an skb still on a
> list (from 00e27108).
> May 19 00:51:36 proxy kernel: Warning: kfree_skb passed an skb still on a
> list (from 00e27304).
> May 19 00:51:36 proxy kernel: general protection: 0000
> May 19 00:51:36 proxy kernel: CPU: 0
> May 19 00:51:37 proxy kernel: EIP: 0010:[<00000003>]
> May 19 00:51:37 proxy kernel: EFLAGS: 00010202
> May 19 00:51:37 proxy kernel: eax: 00000001 ebx: 00380098 ecx:
> 00e2720c edx: 00e27000
> May 19 00:51:37 proxy kernel: esi: 00000184 edi: 00000000 ebp:
> 00e27260 esp: 0019dd30
> May 19 00:51:37 proxy kernel: ds: 0018 es: 0018 fs: 002b gs: 0018
> ss: 0018
> May 19 00:51:37 proxy kernel: Process swapper (pid: 0, process nr: 0,
> stackpage=0019bee8)
> May 19 00:51:37 proxy kernel: Stack: 00137103 00380098 00380098 00e27304
> 00137403 00380098 00e27304 ffffffc4
> May 19 00:51:37 proxy kernel: 00e69ad8 00e27304 00140248 00e27304
> 00000000 00e69ad8 00000000 00e69ad8
> May 19 00:51:37 proxy kernel: 00e69c58 000000e0 000000e0 00e27260
> 00e27304 00140631 00e69ad8 0088aeac
> May 19 00:51:37 proxy kernel: Call Trace: [sock_wfree+35/44]
> [kfree_skb+183/244] [ip_glue+308/472] [ip_defrag+837/860]
> [ip_rcv+380/1464] [def_callback2+38/44] [packet_rcv+177/188]
> May 19 00:51:37 proxy kernel: [net_bh+252/284]
> [do_bottom_half+59/96] [handle_bottom_half+11/32] [sys_idle+92/112]
> [system_call+85/128] [init+0/612] [start_kernel+449/460]
> May 19 00:51:37 proxy kernel: Code: 00 6f ef 00 f0 c3 e2 00 f0 6f ef 00 f0
> 6f ef 00 f0 54 ff 00
> May 19 00:51:37 proxy kernel: Aiee, killing interrupt handler
> May 19 00:51:37 proxy kernel: kfree of non-kmalloced memory: 0019df30,
> next= 00000000, order=0
> May 19 00:51:37 proxy kernel: kfree of non-kmalloced memory: 0019df20,
> next= 00000000, order=0
> May 19 00:51:37 proxy kernel: kfree of non-kmalloced memory: 0019e434,
> next= 00000000, order=0
> May 19 00:51:37 proxy kernel: idle task may not sleep
> May 19 00:51:37 proxy last message repeated 4 times
> May 19 00:51:37 proxy kernel: Warning: kfree_skb passed an skb still on a
> list (from 0088aeec).
> May 19 00:51:37 proxy kernel: Unable to handle kernel paging request at
> virtual
> address eb1e0008
> May 19 00:51:37 proxy kernel: current->tss.cr3 = 00101000, Xr3 = 00101000
> May 19 00:51:37 proxy kernel: *pde = 00000000
> May 19 00:51:37 proxy kernel: Oops: 0000
> May 19 00:51:37 proxy kernel: CPU: 0
> May 19 00:51:37 proxy kernel: EIP: 0010:[<2b1e0008>]
> May 19 00:51:37 proxy kernel: EFLAGS: 00010202
> May 19 00:51:37 proxy kernel: eax: 2b1e0008 ebx: 0085008f ecx:
> 0088adf4 edx: 0088a000
> May 19 00:51:37 proxy kernel: esi: 00000184 edi: 00000000 ebp:
> 0019dc20 esp: 0019dbb8
> May 19 00:51:37 proxy kernel: ds: 0018 es: 0018 fs: 002b gs: 0000
> ss: 0018
> May 19 00:51:37 proxy kernel: Process swapper (pid: 0, process nr: 0,
> stackpage=0019bee8)
> May 19 06:54:23 proxy syslogd 1.3-0: restart.
> May 19 06:54:25 proxy atd[101]: Removing stale lockfile for pid 101
> May 19 06:54:25 proxy /usr/sbin/cron[104]: (CRON) STARTUP (fork ok)
> May 19 06:54:25 proxy kernel: klogd 1.3-0, log source = /proc/kmsg
> started.
> May 19 06:54:26 proxy kernel: Loaded 3057 symbols from /System.map.
> May 19 06:54:26 proxy kernel: Symbols match kernel version.
>
> again, I have no clue about the above. It's even possible that Squid isn't
> responsible for the hangup - though there are only the
> 'standard-processes' running (no FTP, TELNET, MAIL etc.)
>
> Also, there's no cron planned for that timespace.
>
> thanks, sascha

-- 
-----BEGIN GEEK CODE BLOCK-----
Version: 3.1
GAT d- s++: a C++++$ UL++++B+++S+++C++H++U++V+++$ P+++$ L+++ E-
W+++(--)$ N++ w++$>--- t+ 5++ X+() R+ tv b++++ DI+++ e- h-@ 
------END GEEK CODE BLOCK------