| Message ID | 71604351584F6A4EBAE558C676F37CA417BD78AA@H3CMLB02-EX.srv.huawei-3com.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
A better fix is to _not_ disconnect on o2net timeout once a connection has been cleanly established. Only disconnect on o2hb timeout. The reconnects are a problem as we could lose packets and not be aware of it leading to o2dlm hangs. IOW, this patch looks to be papering over one specific problem and does not fix the underlying issue. On Tue, May 7, 2013 at 7:43 PM, Guozhonghua <guozhonghua@h3c.com> wrote: > > > Hi, everyone, > > I had have a test with eight nodes and find one issue. > > > The Linux kernel version is 3.2.40. > > > > As I migrate processes from one node to another, those processes is open > the files on the OCFS2 storage. Sometime one node shutdown TCP connection > with that node whose node number is larger because long time without any > message from it. > > As the TCP connection shutdown, the node whose number larger did not > restart connection to the node, whose number is little and shutdown the TCP > connection. > > So I review the code of the cluster and find it may be a bug. > > > > I changed it and have a test. > > > > Is there anybody having time to view and make sure that those changes is > correct? > > Thanks a lot. > > > > The diff file is as below, of the file is /cluster/tcp.c: > > > > root@gzh-dev:/home/dev/test_replace/ocfs2_ko# diff -pu > ocfs2-ko-3.2-compare/cluster/tcp.c ocfs2-ko-3.2/cluster/tcp.c > > --- ocfs2-ko-3.2-compare/cluster/tcp.c 2012-10-29 19:33:19.534200000 +0800 > > +++ ocfs2-ko-3.2/cluster/tcp.c 2013-05-08 09:33:16.386277310 +0800 > > @@ -1699,6 +1698,10 @@ static void o2net_start_connect(struct w > > if (ret == -EINPROGRESS) > > ret = 0; > > + /** Reset the timeout with 0 to avoid connection again */ > > + if (ret == 0) { > > + atomic_set(&nn->nn_timeout, 0); > > + } > > out: > > if (ret) { > > printk(KERN_NOTICE "o2net: Connect attempt to " SC_NODEF_FMT > > @@ -1725,6 +1728,11 @@ static void o2net_connect_expired(struct > > spin_lock(&nn->nn_lock); > > if (!nn->nn_sc_valid) { > > + /** trigger reconnect with other nodes whose node number is > little than local > > + * while they are still able to access the storage > > + */ > > + atomic_set(&nn->nn_timeout, 1); > > + > > printk(KERN_NOTICE "o2net: No connection established with " > > "node %u after %u.%u seconds, giving up.\n", > > o2net_num_from_nn(nn), > > ------------------------------------------------------------------------------------------------------------------------------------- > ???????????????????????????????????????? > ???????????????????????????????????????? > ???????????????????????????????????????? > ??? > This e-mail and its attachments contain confidential information from H3C, > which is > intended only for the person or entity whose address is listed above. Any > use of the > information contained herein in any way (including, but not limited to, > total or partial > disclosure, reproduction, or dissemination) by persons other than the > intended > recipient(s) is prohibited. If you receive this e-mail in error, please > notify the sender > by phone or email immediately and delete it! > > _______________________________________________ > Ocfs2-devel mailing list > Ocfs2-devel@oss.oracle.com > https://oss.oracle.com/mailman/listinfo/ocfs2-devel >
Resending as my reply bounced. On Thu, May 9, 2013 at 10:01 AM, Sunil Mushran <sunil.mushran@gmail.com>wrote: > A better fix is to _not_ disconnect on o2net timeout once a connection has > been > cleanly established. Only disconnect on o2hb timeout. > > The reconnects are a problem as we could lose packets and not be aware of > it > leading to o2dlm hangs. > > IOW, this patch looks to be papering over one specific problem and does > not fix the > underlying issue. > > > > On Tue, May 7, 2013 at 7:43 PM, Guozhonghua <guozhonghua@h3c.com> wrote: > >> >> >> Hi, everyone, >> >> I had have a test with eight nodes and find one issue. >> >> >> The Linux kernel version is 3.2.40. >> >> >> >> As I migrate processes from one node to another, those processes is open >> the files on the OCFS2 storage. Sometime one node shutdown TCP connection >> with that node whose node number is larger because long time without any >> message from it. >> >> As the TCP connection shutdown, the node whose number larger did not >> restart connection to the node, whose number is little and shutdown the TCP >> connection. >> >> So I review the code of the cluster and find it may be a bug. >> >> >> >> I changed it and have a test. >> >> >> >> Is there anybody having time to view and make sure that those changes is >> correct? >> >> Thanks a lot. >> >> >> >> The diff file is as below, of the file is /cluster/tcp.c: >> >> >> >> root@gzh-dev:/home/dev/test_replace/ocfs2_ko# diff -pu >> ocfs2-ko-3.2-compare/cluster/tcp.c ocfs2-ko-3.2/cluster/tcp.c >> >> --- ocfs2-ko-3.2-compare/cluster/tcp.c 2012-10-29 19:33:19.534200000 >> +0800 >> >> +++ ocfs2-ko-3.2/cluster/tcp.c 2013-05-08 09:33:16.386277310 +0800 >> >> @@ -1699,6 +1698,10 @@ static void o2net_start_connect(struct w >> >> if (ret == -EINPROGRESS) >> >> ret = 0; >> >> + /** Reset the timeout with 0 to avoid connection again */ >> >> + if (ret == 0) { >> >> + atomic_set(&nn->nn_timeout, 0); >> >> + } >> >> out: >> >> if (ret) { >> >> printk(KERN_NOTICE "o2net: Connect attempt to " SC_NODEF_FMT >> >> @@ -1725,6 +1728,11 @@ static void o2net_connect_expired(struct >> >> spin_lock(&nn->nn_lock); >> >> if (!nn->nn_sc_valid) { >> >> + /** trigger reconnect with other nodes whose node number >> is little than local >> >> + * while they are still able to access the storage >> >> + */ >> >> + atomic_set(&nn->nn_timeout, 1); >> >> + >> >> printk(KERN_NOTICE "o2net: No connection established with " >> >> "node %u after %u.%u seconds, giving up.\n", >> >> o2net_num_from_nn(nn), >> >>
--- ocfs2-ko-3.2-compare/cluster/tcp.c 2012-10-29 19:33:19.534200000 +0800 +++ ocfs2-ko-3.2/cluster/tcp.c 2013-05-08 09:33:16.386277310 +0800 @@ -1699,6 +1698,10 @@ static void o2net_start_connect(struct w if (ret == -EINPROGRESS) ret = 0; + /** Reset the timeout with 0 to avoid connection again */ + if (ret == 0) { + atomic_set(&nn->nn_timeout, 0); + } out: if (ret) { printk(KERN_NOTICE "o2net: Connect attempt to " SC_NODEF_FMT @@ -1725,6 +1728,11 @@ static void o2net_connect_expired(struct
Hi, everyone, I had have a test with eight nodes and find one issue. The Linux kernel version is 3.2.40. As I migrate processes from one node to another, those processes is open the files on the OCFS2 storage. Sometime one node shutdown TCP connection with that node whose node number is larger because long time without any message from it. As the TCP connection shutdown, the node whose number larger did not restart connection to the node, whose number is little and shutdown the TCP connection. So I review the code of the cluster and find it may be a bug. I changed it and have a test. Is there anybody having time to view and make sure that those changes is correct? Thanks a lot. The diff file is as below, of the file is /cluster/tcp.c: root@gzh-dev:/home/dev/test_replace/ocfs2_ko# diff -pu ocfs2-ko-3.2-compare/cluster/tcp.c ocfs2-ko-3.2/cluster/tcp.c spin_lock(&nn->nn_lock); if (!nn->nn_sc_valid) { + /** trigger reconnect with other nodes whose node number is little than local + * while they are still able to access the storage + */ + atomic_set(&nn->nn_timeout, 1); + printk(KERN_NOTICE "o2net: No connection established with " "node %u after %u.%u seconds, giving up.\n", o2net_num_from_nn(nn), ------------------------------------------------------------------------------------------------------------------------------------- ???????????????????????????????????????? ???????????????????????????????????????? ???????????????????????????????????????? ??? This e-mail and its attachments contain confidential information from H3C, which is intended only for the person or entity whose address is listed above. Any use of the information contained herein in any way (including, but not limited to, total or partial disclosure, reproduction, or dissemination) by persons other than the intended recipient(s) is prohibited. If you receive this e-mail in error, please notify the sender by phone or email immediately and delete it!