OT: VM killing sec relevant daemons
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Dear list, I know I am (slightly) off topc with this question, but for me it is definitely security related. Since a few days the server of one of my customers keeps telling me (in /var/log/messages): prime kernel: __alloc_pages: 0-order allocation failed (gfp=0x1d2/0) prime kernel: VM: killing process sendmail prime kernel: __alloc_pages: 0-order allocation failed (gfp=0x1d2/0) prime kernel: VM: killing process spamassassin several times in a row. After that it starts killing services and daemons, starting (always) with spamd and amavisd, followed by nfsserver and sendmail, with cups only a short time behind. Also the netfilter tables get flushed at some point, but without so much as a hint to when. Does anybody have a clue what this could possibly be? BTW memtest showed absolutely nothing so far... About the machine: P4 2GHz with 256 MB Memory, 2 HDD RAID1 as /, 2 HDD RAID1 as data storage Linux 9.0 Pro with only samba, nfsserver, sendmail, spamd, amavisd and sshd running. TIA Wolfgang - -- - ----------------------------------------------------- Wolfgang Leithner Pinguin-Systeme.at Bereich Systeme und Security EMail: wolfgang.leithner@pinguin-systeme.at http://www.pinguin-systeme.at - ----------------------------------------------------- GPG Key Fingerprint: 21FE FB64 BD83 8385 364A E927 BB2F F331 84FD 12A9 - ----------------------------------------------------- Diese Nachricht wurde elektronisch erstellt und traegt daher weder Unterschrift noch Siegel This message was electronically created and therefore bears neither signature nor seal - ----------------------------------------------------- -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.6 (GNU/Linux) Comment: For info see http://www.gnupg.org iD8DBQFBCkc6uy/zMYT9EqkRApXkAJwK7bf3Y8cf3Slrh5SLcjHXIutYNACgimne dKo6THcJBu199nMqORPXaTk= =YNF1 -----END PGP SIGNATURE-----
Moin Wolfgang! not really a security topic, but, you are essentially running out of virtual memory and the OOM killer is reaping processes to free memory. you can increase memory and swap and watch to see if there is a process leaking memory. what kernel are you running? -- michael Wolfgang Leithner schrieb am Freitag, den 30. Juli 2004:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Dear list, I know I am (slightly) off topc with this question, but for me it is definitely security related.
Since a few days the server of one of my customers keeps telling me (in /var/log/messages): prime kernel: __alloc_pages: 0-order allocation failed (gfp=0x1d2/0) prime kernel: VM: killing process sendmail prime kernel: __alloc_pages: 0-order allocation failed (gfp=0x1d2/0) prime kernel: VM: killing process spamassassin
several times in a row. After that it starts killing services and daemons, starting (always) with spamd and amavisd, followed by nfsserver and sendmail, with cups only a short time behind.
Also the netfilter tables get flushed at some point, but without so much as a hint to when.
Does anybody have a clue what this could possibly be? BTW memtest showed absolutely nothing so far...
About the machine: P4 2GHz with 256 MB Memory, 2 HDD RAID1 as /, 2 HDD RAID1 as data storage Linux 9.0 Pro with only samba, nfsserver, sendmail, spamd, amavisd and sshd running.
TIA
Wolfgang
- -- - ----------------------------------------------------- Wolfgang Leithner Pinguin-Systeme.at Bereich Systeme und Security EMail: wolfgang.leithner@pinguin-systeme.at http://www.pinguin-systeme.at - ----------------------------------------------------- GPG Key Fingerprint: 21FE FB64 BD83 8385 364A E927 BB2F F331 84FD 12A9 - ----------------------------------------------------- Diese Nachricht wurde elektronisch erstellt und traegt daher weder Unterschrift noch Siegel This message was electronically created and therefore bears neither signature nor seal - -----------------------------------------------------
-----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.6 (GNU/Linux) Comment: For info see http://www.gnupg.org
iD8DBQFBCkc6uy/zMYT9EqkRApXkAJwK7bf3Y8cf3Slrh5SLcjHXIutYNACgimne dKo6THcJBu199nMqORPXaTk= =YNF1 -----END PGP SIGNATURE-----
-- Check the headers for your unsubscription address For additional commands, e-mail: suse-security-help@suse.com Security-related bug reports go to security@suse.de, not here
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi Michael!
not really a security topic, I know :( but, you are essentially running out of virtual memory and the OOM killer is reaping processes to free memory. you can increase memory and swap and watch to see if there is a process leaking memory. what kernel are you running? 2.4.21-231-default
There is more than enough swap but the memory is definitely tight. Have to talk to my customer, though. Thanks for the fast answer. Regards Wolfgang -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.6 (GNU/Linux) Comment: For info see http://www.gnupg.org iD8DBQFBClrZuy/zMYT9EqkRAvJxAJ9GViVUktpWyLuEYuGvOhtSnnBKiACeI2H9 /plWgH1v2f2+s0VuFz2x0zQ= =dz/6 -----END PGP SIGNATURE-----
participants (2)
-
Michael Galloway -
Wolfgang Leithner