<div dir="ltr">Hi, <div><br></div><div>I have been running sheepdog on a cluster of three nodes for some time now. I sometime experience errors on the guest and host sides. </div><div><br></div><div>On the guest side I read: </div>
<div>[TIMESTAMP] Buffer I/O error on device vda1, logical block XXXXXX</div><div>...</div><div>[TIMESTAMP] end_request: I/O error, dev vda, sector XXXXXX<br></div><div><br></div><div>On the host side (qemu log) I read: </div>
<div>qemu-system-x86_64: shrinking is not supported</div><div><br></div><div>If I'm not mistaken, this message is writen in the block/sheepdog.c file in qemu: </div><div><br></div><div><div>static int sd_truncate(BlockDriverState *bs, int64_t offset)</div>
<div>{</div><div> BDRVSheepdogState *s = bs->opaque;</div><div> int ret, fd;</div><div> unsigned int datalen;</div><div><br></div><div> if (offset < s->inode.vdi_size) {</div><div> error_report("shrinking is not supported");</div>
<div> return -EINVAL;</div><div> } else if (offset > SD_MAX_VDI_SIZE) {</div><div> error_report("too big image size");</div><div> return -EINVAL;</div><div> }</div><div><br></div><div>
fd = connect_to_sdog(s);</div><div> if (fd < 0) {</div><div> return fd;</div><div> }</div></div><div>...</div><div>}</div><div><br></div><div>I tried to find the call graph for this function but I am no C/C++ developer.</div>
<div><br></div><div>Anyone have had this problem?</div><div><br></div><div>Hadrien Kohl</div><div> </div></div>