Camelia, the Perl 6 bug

IRC log for #gluster-dev, 2013-07-30

| Channels | #gluster-dev index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:52 bala joined #gluster-dev
01:33 asias joined #gluster-dev
01:57 badone_ joined #gluster-dev
02:57 kshlm joined #gluster-dev
03:05 bharata joined #gluster-dev
03:38 shubhendu joined #gluster-dev
03:56 aravindavk joined #gluster-dev
04:01 itisravi joined #gluster-dev
04:58 lalatenduM joined #gluster-dev
05:04 kanagaraj joined #gluster-dev
05:06 raghu joined #gluster-dev
05:07 bulde1 joined #gluster-dev
05:18 lalatenduM joined #gluster-dev
05:37 bala joined #gluster-dev
05:42 bharata-rao asias, hi
05:42 asias bharata-rao, hi
05:43 bharata-rao asias, I wasn't able to recreate your qemu-img create problem at all with latest gluster and qemu
05:43 bharata-rao asias, I understand the problem which stefanha described, but somehow it is not affecting qcwo2 on gluster
05:43 asias bharata-rao, right, upstream qemu is ok
05:44 asias bharata-rao, rhel6 is not right
05:45 bharata-rao asias, but upstream qemu assumes that read of 1st sector (512) from gluster backend returns zero
05:45 bharata-rao asias, isn't that the basic problem ?
05:46 asias bharata-rao, yes, that's right. we are lucky in upsteam code ;-)
05:47 bharata-rao asias, right, so even with qcow2 header + remaining junk in the 1st 512 bytes of qcow2 file, things work with upstream qemu but not on rhel6
05:47 asias bharata-rao right
05:47 bharata-rao asias, interesting
05:48 asias bharata-rao, I even tried memset 0xff to the 512 buffer, upstream code still works
05:49 asias bharata-rao, did you see Morita's patch?
05:50 bharata-rao asias, so that means rhel6 qemu is trying to interpret the data beyond the qcow2 header in the 1st 512 byte sector, but upstream qemu is not ?
05:50 bharata-rao asias, I did see, not in detail though
05:51 asias bharata-rao, on rhel6, qcow2_open:qcow2_read_extensions fails with qemu-img create -f qcow2 gluster://...
05:52 asias bharata-rao, there are qcow2 extensions after the qcow2 header
05:53 bharata-rao asias, oh ok
05:53 asias bharata-rao, the header is terminated by QCOW2_EXT_MAGIC_END magic which is 0
05:53 asias bharata-rao, with remaing junk after the header, we get Qcow2UnknownHeaderExtension
05:54 asias bharata-rao, then it tried to read a random file open set in "ret = bdrv_pread(bs->file, offset , uext->data, uext->len);"
05:57 bharata-rao asias, ok, but that still doesn't explain why it is working with upstream qemu since there too the header extensions follow the header right ?
05:58 asias bharata-rao, right! this is the interesting part
06:01 wushudoin joined #gluster-dev
06:01 mohankumar joined #gluster-dev
06:14 ndarshan joined #gluster-dev
07:05 vshankar joined #gluster-dev
07:30 badone_ joined #gluster-dev
08:03 mohankumar joined #gluster-dev
08:33 puebele joined #gluster-dev
08:51 puebele1 joined #gluster-dev
08:53 ndarshan joined #gluster-dev
08:55 itisravi_ joined #gluster-dev
09:01 itisravi joined #gluster-dev
09:02 bharata joined #gluster-dev
09:07 deepakcs joined #gluster-dev
09:56 kkeithley1 joined #gluster-dev
10:14 kshlm joined #gluster-dev
10:19 kkeithley joined #gluster-dev
10:20 kkeithley joined #gluster-dev
10:23 kkeithley joined #gluster-dev
10:27 edward1 joined #gluster-dev
10:28 kkeithley joined #gluster-dev
11:20 ndarshan joined #gluster-dev
13:19 kkeithley joined #gluster-dev
13:40 lalatenduM joined #gluster-dev
13:58 Technicool joined #gluster-dev
14:22 lalatenduM joined #gluster-dev
14:30 wushudoin joined #gluster-dev
14:37 kkeithley joined #gluster-dev
15:02 lpabon joined #gluster-dev
15:25 jclift joined #gluster-dev
16:11 bala joined #gluster-dev
16:33 xavih joined #gluster-dev
17:07 Technicool joined #gluster-dev
17:10 bulde joined #gluster-dev
18:54 raghu` joined #gluster-dev
19:57 bulde joined #gluster-dev

| Channels | #gluster-dev index | Today | | Search | Google Search | Plain-Text | summary