re-creation of __db.001, __db.002, __db.003
Iosif Fettich <[email protected]>
| Newsgroups | gmane.linux.rpm.yum |
|---|---|
| Message-ID | <[email protected]> |
Dear fellow yummers, I'd be grateful for any hint regarding a small but apparently weird issue. Taking care of two very similar servers, I've noticed a difference in behavior that I would like to understand, but where my work so far didn't bring any light onto the matter. When running /usr/bin/yum --downloadonly update -y on them, as root, server A _always_ recreates /var/lib/rpm/__db.00[1-3], whereas server B _never_ does this, regardless if there are updates to be downloaded or not. Both servers run CentOS release 5.5 (Final), on both servers the installed versions for yum and rpm are the same, yum-metadata-parser-1.1.2-3.el5.centos yum-updatesd-0.9-2.el5 yum-fastestmirror-1.1.16-14.el5.centos.1 yum-downloadonly-1.1.16-14.el5.centos.1 yum-priorities-1.1.16-14.el5.centos.1 yum-3.2.22-26.el5.centos yum-utils-1.1.16-14.el5.centos.1 rpm-python-4.4.2.3-18.el5 rpm-4.4.2.3-18.el5 rpmforge-release-0.5.1-1.el5.rf rpm-libs-4.4.2.3-18.el5 The config files on both servers seem to be identical, they also behave OK both - when there are upgrades, both servers see them, both do install them, I cannot complain about any sign of mis-behavior on any of the servers. Still, whereas the one _always_ re-creates the RPM lock files - i.e., I see a different modification time on them, the other one _never_ does so - i.e., the exposed modification time never changes, when running yum as shown above. This is puzzling, and I'm really anxious to learn what's causing this difference in behavior. I've strace-ed the process on both servers, but the output is rather huge and I wasn't able yet to spot any significant difference between them. Kernel versions are different, but vanilla - 2.6.18-194.3.1.el5 on A, 2.6.18-92.1.10.el5 on B (but upgrading the kernel on B isn't possible right now, neighter downgrading the kernel on A). Many thanks in advance for any idea! Iosif Fettich