Messages in this thread | | | Date | Thu, 12 Aug 1999 12:44:53 -0400 | From | Ben McCann <> | Subject | Re: 2.2.11: Complicated memory leak... |
| |
I've been off the list for a week so excuse me if this repeats what others have reported.
I have two machines which both suffer catastrophic, and sudden, memory leaks.
M1 = Pentium MMX /64MB, IDE, 3c59x NIC (100 BT) M2 = Pentium II /128MB, IDE, Intel EtherExpress Pro (NIC) (100BT)
Both machines are NFS clients that primarily access Solaris servers. They are both used for SW devlopment with no games, MP3, etc...
They fail in one of two ways:
1. If I'm in X then its likely to totally lock up. However, I have had a couple failures where I could switch back to the text console and use Alt-SysRq as described next...
2. I've also had the network interface STOP working after a 1 second pause. (When I say 'stop', I mean I couldn't even ping). I could recover the interface by doing a 'down' and then 'up' with 'ifconfig'.
Getting back to the memory leak... After I noticed the interface was down, I observed that Alt-SysRq-M shows about 150000 packet buffers allocated on my 128MB machine with basically no free memory left. (Really there was 3 or 4 MB).
I've checked that every half-hour or so and it appears to happen all at once. I go from 50 buffers to 150000!
I've hacked 'alloc_skb' on one machine to panic if the number of allocated sk_buff's exceeds 1000. I'll try to get a meaning stack trace and post it to the list. Meanwhile, its back to 2.2.10 on the other PC.
-Ben McCann
-- Ben McCann Indus River Networks 31 Nagog Park Acton, MA, 01720 email: bmccann@indusriver.com web: www.indusriver.com phone: (978) 266-8140 fax: (978) 266-8111
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |